Field Assembly LLC
What’s The Diff?
A dated record of what the AI companies' published rules said before they changed.
Read it newest first below, by operator, or by kind of change.
Why we keep it
The companies building AI systems publish the rules for using them: terms of service, usage policies, privacy policies, notices about which models are going away. Those documents change, and when one does the previous wording usually goes with it. The page is replaced, and the version you agreed to, or built on, or quoted in a filing, is no longer anywhere you can point to, because the company holds the only copy. This page is that copy, kept by someone else.
Field Assembly keeps a weekly, externally timestamped capture of the published governance surface of the major AI operators: their consumer terms, usage and acceptable-use policies, privacy policies, model deprecation notices, and capability and safety claims. When a tracked document changes, an entry appears here: what the document said, what it says now, both quoted verbatim from retained captures, with the dates.
Entries state changes; they interpret nothing and take no position on any vendor. Detection is automatic. Every entry is reviewed before publication, and the captures behind it are hash-chained and externally timestamped under the record standard, so what a document said, on a date, can be proven to a third party later.
What the docket watches
62 documents across 17 operators. Capture began 2026-08-28; each operator is listed with the date its own coverage starts.
Every operator in the docket appears here whether or not its documents have changed, because a tracked document that has not changed is a finding too. Where a document could not be reached, that is stated rather than counted as unchanged.
- Amazon — 7 entries · 4 documents tracked; tracked since 2026-08-30
- Anthropic — 17 entries · 6 documents tracked; tracked since 2026-08-28
- Apple — 4 entries · 4 documents tracked; tracked since 2026-08-30
- ByteDance — 0 entries · 2 documents tracked; no changes found; tracked since 2026-08-30
- Canva — 8 entries · 6 documents tracked; tracked since 2026-08-30
- Character.AI — 1 entry · 1 document tracked; tracked since 2026-08-30
- Google — 6 entries · 5 documents tracked; tracked since 2026-08-28
- Meta — 1 entry · 2 documents tracked; tracked since 2026-08-29
- Microsoft — 2 entries · 2 documents tracked; tracked since 2026-08-28
- OpenAI — 18 entries · 9 documents tracked; tracked since 2026-08-28
- Perplexity — 3 entries · 2 documents tracked; tracked since 2026-08-28
- Replika — 2 entries · 2 documents tracked; tracked since 2026-08-30
- Roblox — 5 entries · 3 documents tracked; tracked since 2026-08-30
- Samsung — 1 entry · 3 documents tracked; tracked since 2026-08-30
- Snap — 7 entries · 5 documents tracked; tracked since 2026-08-30
- xAI — 8 entries · 3 documents tracked; tracked since 2026-08-28
- Zoom — 1 entry · 3 documents tracked; tracked since 2026-08-30
The same changes grouped by what structurally changed in the document — scope, responsibility, a withdrawn commitment, a deleted notice — are indexed at Kinds of change. The grouping is a finding aid; the entries below are the record.
Show entries from
2026-11-21 · snap — Snap adds an introductory "Welcome" section to its Privacy Policy
Snap Privacy Policy · detected weekly pass, 2026-11-21
Before:
prior Privacy Policy, which applies to all users until September 21, 2026, here . Company Snap Inc. Careers News Privacy and Safety Community Snapchat
After:
prior Privacy Policy, which applies to all users until September 21, 2026, here . Privacy Policy Effective: September 21, 2026 Welcome to Snap Inc. ’s Privacy Policy. This Policy explains how we collect and use your data and how you can control your information. Looking for a quick summary on our privacy practices? Check out this page or this video . If you’re looking for specific product privacy information, for example, how we process your Chats and Snaps, take a look at our Privacy by Product page. In addition to these documents, we also show in-app notices that provide you more information about our products and services. If you’re a SPECS user you can learn more on how we collect and use your data here . Transparency is one of our core values at Snap. We believe there shouldn’t be any surprises about the data we collect and how we use it — that’s why we’re upfront with how we process it. For example, we process your information to provide you a more personalized experience, including to show you content and information that is most relevant to your experience, as well as more relevant advertisements. Understanding your interests and preferences help us provide a better product experience. Furthermore, we believe a personalized experience should not come at the expense of your privacy. Just like in real life, there are some moments you want to share privately with your close friends, and others that you want to share publicly. That’s why from day one our philosophy has be
A new introductory section states the Policy's purpose, points to a summary page and video and to the Privacy by Product page, and describes Snap's stated approach to personalization and privacy, where none of that text appeared before. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · roblox — Roblox removes its Terms of Use "summary of changes" banner and adds two sidebar links
Roblox Terms of Use (contains AI Features section) · detected weekly pass, 2026-11-21
Before:
of the Roblox Terms OSA Annex Advertising Integrations Terms Roblox Accessibility Statement See more Roblox Terms of Use Effective Date: November 1, 2026 We are updating our Terms of Use to make them clearer and to better explain how Roblox works. Here’s a high-level summary of the changes: We have tried to simplify the language and reorganize sections of the Terms of Use, including the User Terms and Creator Terms. We are clarifying Roblox’s role on how user generated content (UGC) is provided to you on the platform. In general, Roblox acts as the principal distributor and merchant of record for purchases made through Roblox. This means that when you access or purchase any content on the Roblox Platform, you are obtaining the content from Roblox. Users and Creators do not enter into a direct transaction with each other for those purchases. We are also clarifying related sections on Robux, payments, tax, and Creator payment. For Users in the EEA only, we added the right of withdrawal for consumers. USER TERMS CREATOR TERMS Appendix A (Japan) Appendix B (South Korea) Appendix
After:
of the Roblox Terms OSA Annex Advertising Integrations Terms Roblox Accessibility Statement Age-Restricted Content on Roblox Roblox Subscription Terms See more Roblox Terms of Use Effective Date: November 1, 2026 USER TERMS CREATOR TERMS Appendix A (Japan) Appendix B (South Korea) Appendix
The paragraph summarizing what the November 1, 2026 Terms update changed — simplified organization, Roblox's role as principal distributor/merchant of record for UGC purchases, Robux/payment/tax/Creator-payment clarifications, and an EEA right of withdrawal — is removed from the page, while the effective date stays November 1, 2026. The sidebar now also links to "Age-Restricted Content on Roblox" and "Roblox Subscription Terms." 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · roblox — Roblox removes its Privacy Policy "summary of recent changes" banner and adds two sidebar links
Roblox Privacy and Cookie Policy · detected weekly pass, 2026-11-21
Before:
of the Roblox Terms OSA Annex Advertising Integrations Terms Roblox Accessibility Statement See more Roblox Privacy and Cookie Policy Effective Date: November 1, 2026 Summary of Recent Changes We are updating our Roblox Privacy and Cookie Policy , effective November 1, 2026, to reflect certain changes we are making on Roblox. Here’s a high-level summary of the changes we are making: We added language describing regional variation for access to certain features. We added information about data collection from researchers. We added language describing our practices for sharing information with authorities. You can also find the full Privacy Policy here . Welcome to
After:
of the Roblox Terms OSA Annex Advertising Integrations Terms Roblox Accessibility Statement Age-Restricted Content on Roblox Roblox Subscription Terms See more Roblox Privacy and Cookie Policy Effective Date: November 1, 2026 You can also find the full Privacy Policy here . Welcome to
The paragraph summarizing what the November 1, 2026 update changed — regional variation in feature access, data collection from researchers, and practices for sharing information with authorities — is removed from the page, while the effective date stays November 1, 2026. The sidebar now also links to "Age-Restricted Content on Roblox" and "Roblox Subscription Terms." 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · openai — OpenAI Safety page replaces hub links with a dated research-post index
OpenAI Safety page · detected weekly pass, 2026-11-21
Before:
on safety Explore the deployment safety hub (opens in a new window) GPT‑5.3‑Codex System Card Learn more Addendum to GPT‑5.2 System Card: GPT‑5.2‑Codex Learn more Update to GPT‑5
After:
what it has learned about strengthening safeguards. (opens in a new window) Nov 18, 2026 Managing frontier AI risk: From training to deployment How OpenAI evaluates and manages frontier AI risk across research, internal deployment, and external deployment. (opens in a new window) Nov 12, 2026 Improving Fuzzy Robustness in LLM Safety with Automated Red Teaming This research evaluates whether automated red teaming can strengthen model safeguards in settings where acceptable use depends on context. Nov 10, 2026 What affects chain-of-thought monitorability across the model lifecycle? This research studies how training and architecture affect whether a model’s chain of thought is useful for oversight. Nov 3, 2026 Eval-production behavior gaps in agent safety and the role of eval-awareness We study agent behavior in safety evaluations and in production, and explore how models' interpretations of evaluation objectives relate to eval-production performance gaps. Nov 2, 2026 Training activation classifiers for safety and alignment We show how we train and deploy efficient monitors that use a model’s internal activity to help detect and interrupt policy-violating behavior in real time. Oct 30, 2026 Community safety report and new protections for teens Updates to violence prevention, safeguards for younger users, and plans for recurring community safety reporting. (opens in a new window) Oct 29, 2026 How we detect frontier model misalignment How we monitor frontier models, investigate c
The single changed span replaces System Card and safety-hub links with a list of dated safety research posts, running from Oct 29 through Nov 18, 2026. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · openai — OpenAI adds deprecation entry for GPT-5.2/5.4 EU fast-tier models
OpenAI API deprecations · detected weekly pass, 2026-11-21
Before:
fine-tuning job creation Restricted access See the fine-tuning timeline . May 7, 2026 Feb 26, 2027 /v1/audio/translations Deprecated /v1/realtime/translations Aug 26, 2026 Model deprecation
After:
fine-tuning job creation Restricted access See the fine-tuning timeline . May 7, 2026 Jan 11, 2027 GPT-5.2, GPT-5.4, and GPT-5.4 mini with EU data residency and service_tier: fast or priority Deprecated The same models with EU data residency and Standard processing ( service_tier: default ), or the GPT-6.2 model family . Nov 18, 2026 Feb 26, 2027 /v1/audio/translations Deprecated /v1/realtime/translations Aug 26, 2026 Model deprecation
A new entry states that GPT-5.2, GPT-5.4, and GPT-5.4 mini with EU data residency and service_tier fast or priority are deprecated as of Nov 18, 2026 and retire Feb 26, 2027, with the Standard-processing versions or the GPT-6.2 family named as the recommended replacement. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · google — Google renames the Gemini History setting and expands what it covers
Gemini Apps Privacy Hub · detected weekly pass, 2026-11-21
Before:
you know about me?". How can I control whether Google uses my audio and Gemini Live videos & screenshares to improve its services? You
After:
you know about me?". How can I control whether Google uses my chat content, including text & attachments, to train generative AI models? You can control whether Google uses your chat content, including text & attachments, to train generative AI models in Gemini History. If this setting is on: Google uses the content of your future chats to train generative AI models. This includes: What you say to Gemini Apps, including your prompts and transcripts of audio What you share with Gemini Apps, like photos, files, audio or video attachments , and screens you ask about Gemini Apps’ responses Tasks and tool calls Info about your location Data from apps you connect to Gemini Related product usage data, like your language and device type To protect your privacy, chats are separated from your account before being sent to service providers to help improve generative AI models. Reviewed chats and related data are retained for up to 3 years, even if you delete them or turn this setting off later. If this setting is off: Your future chats won’t be used to train generative AI models, unless you provide feedback . Future chats will be reviewed and used to improve Google services in other ways, and to protect Google, its users, and the public. Chats reviewed by service providers are separated from your account and kept for up to 3 years. This setting applies only to your Gemini history. Turning off this setting doesn’t impact activity saved and managed by other Google services, including acti
The FAQ question and its description are renamed and expanded: "Gemini Live videos & screenshares" becomes "chat content, including text & attachments," and the listed data now includes prompts, transcripts, shared files, tool calls, location, and connected-app data, retained up to 3 years. The hub's own last-updated date moved from October 26 to November 19, 2026, and the companion notice's from October 13 to November 16. 65 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · canva — Canva changes how Team seat charges are billed, effective November 16
Canva Terms of Use · detected weekly pass, 2026-11-21
Before:
(d) below. b. Pricing for Additional Seats on your Team . If you're on a plan that enables you to create a Team, you're billed according to Canva’s “Pay as You Grow” model. When you add seats to a Team, you will be billed for them on your next Billing Date (as defined below) unless they are removed from your Team by the Team Owner or an Administrator prior to the Billing Date. Three (3) days prior to the Billing Date, Canva will notify via email the billing contact associated with your Canva Team account of the number of additional seats and the associated subscription fees. The applicable “Billing Date” is as follows: (i) if you're on a monthly subscription, the date of your next monthly renewal; (ii) if you're on an annual subscription, every 3 months after the start date of your subscription. You will not receive a refund or credit for removing seats from your Team that have already been paid for. The Pay as You Grow model allows you to scale your Team as needed and reconcile costs for anyone who has joined your Team but will no longer need access to the Service before you are billed for them. c. Taxes . Your subscription fees are inclusive of all taxes unless
After:
(d) below. b. Pricing for Additional Seats on your Team . If you’re on a plan that enables you to create a Team, your subscription fees are based on the number of seats in your Team. The Team Owner can add seats to your Team, and authorises Canva to charge the payment method nominated for your Team for any seats they add. A seat is added when a Team Owner invites a person to your Team or adds a seat, whether or not the invitation has been accepted. When a seat is added part-way through a billing period, we’ll charge you for it at the time it’s added. The charge is prorated so you only pay for the remainder of your current billing period: (i) if you’re on a monthly subscription, until your next monthly renewal; (ii) if you’re on an annual subscription, until your next renewal. c. Taxes . Your subscription fees are inclusive of all taxes unless
The Terms' effective date moves from October 30 to November 16, 2026. The Team seat-billing clause changes from billing added seats on the next scheduled Billing Date to charging the Team Owner's payment method at the time a seat is added, prorated to the remainder of the current billing period. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · canva — Canva's policy archive index adds a November 16, 2026 Terms of Use entry
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-11-21
Before:
30, 2025 Tag a Teacher Competition Terms Current Terms of Use Current Oct 29, 2026 Aug 19, 2026 Jul 20, 2026 Jun 23, 2026
After:
30, 2025 Tag a Teacher Competition Terms Current Terms of Use Current Nov 16, 2026 Oct 29, 2026 Aug 19, 2026 Jul 20, 2026 Jun 23, 2026
The archive's Terms of Use date list adds "Nov 16, 2026" ahead of the prior "Oct 29, 2026" entry, consistent with the Terms of Use's new effective date. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-21 · anthropic — Anthropic adds Claude Code guidance to the claude-sonnet-4-5 retirement entry
Anthropic model deprecations · detected weekly pass, 2026-11-21
Before:
API. Retirement date Deprecated model Recommended replacement November 30, 2026 claude-sonnet-4-5-20250929 claude-sonnet-5-5 2026-06-05: Claude Opus 4.1 model This model was retired August 5,
After:
API. Retirement date Deprecated model Recommended replacement November 30, 2026 claude-sonnet-4-5-20250929 claude-sonnet-5-5 This retirement includes claude-sonnet-4-5 , the same model’s ID without a date. If you use Claude Code, see Overview for which dates apply and Migrating to replacements for what to change. Once this model is retired, Claude Code on the Claude API shows an error that begins There's an issue with the selected model . If you set a fallback model chain , Claude Code switches to the next model in the chain instead. It tries Claude Sonnet 4.5 again on your next message. 2026-06-05: Claude Opus 4.1 model This model was retired August 5,
The retirement entry for claude-sonnet-4-5-20250929 (which includes its undated alias claude-sonnet-4-5) now states that Claude Code on the Claude API will show an error once the model retires, and that a configured fallback model chain will switch to the next model and retry Sonnet 4.5 on the following message. 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-14.
2026-11-14 · snap — Snap Terms of Service add a plain-language summary, including new language on AI-enabled features
Snap Terms of Service · detected weekly pass, 2026-11-14
Before:
Service, which apply to all users until September 21, 2026, here . Company Snap Inc. Careers News Privacy and Safety Community Snapchat Support Spectacles
After:
Service, which apply to all users until September 21, 2026, here . Click here for a helpful summary of the Snap Terms of Service The Snap Terms of Service is a contract between you and Snap. It explains the rules for using Snapchat and our other services (which we call our “Services”). These rules are legally binding, so it's important to understand them. Here’s a quick breakdown of the key points, but you should still read the whole thing: You must be 13 (or older, depending on where you live) to use Snapchat : And some of our other Services may require that you are 18 or older. We collect and use your information to provide the Services : See our Privacy Policy for details. Our Services provide a personalized experience : We show ads and recommend content based on your interests as explained here . You are responsible for your content and any activity that occurs under your account : But we can change, limit access to, or remove that content. Don’t use or share anyone else’s content without their permission. And don’t use the Services to create or share any content that violates someone’s rights, including intellectual property or privacy. Our Services include AI-enabled features: These features use the inputs directed by you to generate outputs like content or responses, which may not be accurate or appropriate, so you use them at your own risk. All inputs and anything you generate using the AI features or ask the AI features to generate must comply with the Snap Terms of
The terms added a "quick breakdown" summary near the top covering the age requirement, data collection for providing the Services, personalized recommendations, and user responsibility for content. The summary also states that the Services include AI-enabled features whose outputs "may not be accurate or appropriate," used "at your own risk," and that all AI inputs and outputs must comply with the Snap Terms of Service. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · snap — Snap's "Generative AI on Snapchat" FAQ adds a new entry on Spotlight AI labels
Generative AI on Snapchat · detected weekly pass, 2026-11-14
Before:
you know that the image was created with a generative AI-powered feature. What is the Snap Ghost with sparkles watermark? Some generative AI-powered features,
After:
you know that the image was created with a generative AI-powered feature. On Spotlight, you may also see ‘Creator labeled AI-generated’ when a creator marked a Snap as AI-generated, or when Snap identifies AI-generated content. Learn more in How do AI labels work on Spotlight? How do AI labels work on Spotlight? When you post to Spotlight, you can turn on ‘Add AI Label’ on the New Spotlight Post screen if your content was made with AI. Viewers may then see ‘Creator labeled AI-generated’ on that Snap. An AI label can also appear when Snap identifies AI-generated content. For step-by-step posting help, see How do I post Snaps to Spotlight? For how to add the label, see How do I label AI-generated content on Spotlight ? What is the Snap Ghost with sparkles watermark? Some generative AI-powered features,
The page added a new FAQ entry, "How do AI labels work on Spotlight?", explaining that posters can turn on "Add AI Label" when posting to Spotlight, and that Snap may also apply a "Creator labeled AI-generated" label when it identifies AI-generated content itself. Two existing questions were updated to reference the new entry and a companion article, "How do I label AI-generated content on Spotlight?" 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · roblox — Roblox generative-AI creator rules rename "Content Maturity" to "Content Ratings" throughout
Roblox Games with Generative AI (creator rules) · detected weekly pass, 2026-11-14
Before:
3D generations, and avatar movement, you must disclose it within the Content Maturity questionnaire. While merely including an AI component in your game does not impact the content maturity of your game, games with extended AI interactions must include the appropriate Restricted content maturity label so that they are only available to the appropriate audience. Limited interactions For games without the Restricted content maturity label, model outputs must adhere to the underlying Content Maturity classification of the game. When using Roblox-served AI tools, such as GenerationService
After:
3D generations, and avatar movement, you must disclose it within the Content Ratings Questionnaire . While merely including an AI component in your game does not impact your game's content ratings, games with extended AI interactions must include the appropriate Restricted content maturity rating so that they are only available to the appropriate audience. Limited interactions For games without the Restricted content maturity rating, model outputs must adhere to the underlying content maturity classification of the game. When using Roblox-served AI tools, such as GenerationService
The page consistently renamed its rating terminology: "Content Maturity questionnaire" became "Content Ratings Questionnaire," and "content maturity rating/label" became "content rating" or "content descriptor" in the same sentences. The underlying rules in the listed spans are otherwise unchanged — the same disclosure, Restricted-rating, and Limited-interactions requirements apply, just under the new terms. 4 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · openai — OpenAI Service Terms add "Businesses in ChatGPT," Business Agents, and ChatGPT Wallet terms
OpenAI Service Terms · detected weekly pass, 2026-11-14
Before:
another service will be subject to that service’s terms, policies, and configuration.
After:
another service will be subject to that service’s terms, policies, and configuration. 16. Businesses in ChatGPT Businesses in ChatGPT allows businesses to claim and manage their business listings, configure Business Agents, and access related tools and insights. If you access or use Businesses in ChatGPT, the Businesses in ChatGPT Terms apply.
The terms updated the effective date to November 13, 2026, added a new Section 16 covering "Businesses in ChatGPT" (claiming and managing business listings and configuring "Business Agents"), and extended existing GPT provisions — the endorsement disclaimer, abuse reporting, and creator removal rights — to also cover "Business Agents." The terms also state that ChatGPT Wallet payments are governed by separate ChatGPT Wallet Terms of Service and Privacy Notice, and extended the ChatGPT for Healthcare provision to cover PHI that Customer transmits through the Service. 7 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · openai — OpenAI API deprecations page adds Oct 23 shutdowns for gpt-3.5-turbo and gpt-4-0613 families, restructures lifecycle guidance
OpenAI API deprecations · detected weekly pass, 2026-11-14
Before:
page . Base GPT models Shutdown date Deprecated model Deprecated model price Recommended replacement 2024-01-04 ada $0.40 / 1M tokens babbage-002 2024-01-04 babbage $0.50 / 1M tokens babbage-002 2024-01-04 curie $2.00 / 1M tokens davinci-002 2024-01-04 davinci $20.00 / 1M tokens davinci-002 2024-01-04 code-davinci-002 — gpt-3.5-turbo-instruct Pricing for the replacement babbage-002 and davinci-002 models can be found on the pricing page . Edit models & endpoint Shutdown date Model / system Recommended replacement 2024-01-04 text-davinci-edit-001 gpt-4o 2024-01-04 code-davinci-edit-001 gpt-4o 2024-01-04 /v1/edits /v1/chat/completions Fine-tuning GPT models Shutdown date Deprecated model Training price Usage price Recommended replacement 2024-01-04 ada $0.40 / 1M tokens $1.60 / 1M tokens babbage-002 2024-01-04 babbage $0.60 / 1M tokens $2.40 / 1M tokens babbage-002 2024-01-04 curie $3.00 / 1M tokens $12.00 / 1M tokens davinci-002 2024-01-04 davinci $30.00 / 1M tokens $120.00 / 1K tokens davinci-002 , gpt-3.5-turbo , gpt-4o First-generation text embedding models Shutdown date Deprecated model Deprecated model price Recommended replacement 2024-01-04 text-similarity-ada-001 $4.00 / 1M tokens text-embedding-3-small 2024-01-04 text-search-ada-doc-001 $4.00 / 1M tokens text-embedding-3-small 2024-01-04 text-search-ada-query-001 $4.00 / 1M tokens text-embedding-3-small 2024-01-04 code-search-ada-code-001 $4.00 / 1M tokens text-embedding-3-small 2024-01-04 code-search-ada-text-001 $4.
After:
gpt-3.5-turbo Jun 13, 2023 Jun 13, 2024 Earliest estimate gpt-4-0314 Historical price $30.00 / 1M input tokens + $60.00 / 1M output tokens gpt-4o
The page restructured "Past deprecations" into a table and added entries dated Oct 23, 2026 for gpt-3.5-turbo-0125, gpt-3.5-turbo, gpt-3.5-turbo-completions, gpt-4-0613, gpt-4, and related models, pointing to gpt-5.6-terra as the recommended replacement. It also added definitions for three model lifecycle states (Active, Deprecated, legacy), extended the minimum notice period for generally-available models to also cover image generation models, and removed language reserving the right to retire a model sooner for safety or compliance reasons. 6 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · canva — Canva Policy Archives index adds "Branded Content Terms" and new Nov 11, 2026 versions of two agreements
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-11-14
Before:
2026 Canva for Enterprise - Master Services Agreement Nov 5, 2021 Canva's Content License Agreement Current Apr 14, 2026 Mar 20, 2026 Sep 18, 2025 Feb 4, 2025
After:
2026 Canva for Enterprise - Master Services Agreement Nov 5, 2021 Canva's Branded Content Terms Current Canva's Content License Agreement Current Nov 11, 2026 Apr 14, 2026 Mar 20, 2026 Sep 18, 2025 Feb 4, 2025
The index added a new document, "Canva's Branded Content Terms" (listed as Current), and added a Nov 11, 2026 "Current" version of "Canva's Content License Agreement" above its prior Apr 14, 2026 version. The other listed span shows the same pattern for "Canva's Music License Agreement," which also gained a Nov 11, 2026 Current version listed above its prior Nov 18, 2024 version. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-14 · anthropic — Anthropic adds Claude Fable 5.2 and Mythos 5.2 to the model-status table, renames Bedrock table references
Anthropic model deprecations · detected weekly pass, 2026-11-14
Before:
more details, see Commitments on Model Deprecation and Preservation . Model status Current and recently retired models are listed in the following table with their status: API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-mythos-5-1 Active N/A
After:
more details, see Commitments on Model Deprecation and Preservation . Model status Current and recently retired models are listed in the following table with their status: API model name Current state Deprecated Tentative retirement date claude-fable-5-2 Active N/A Not sooner than November 9, 2027 claude-mythos-5-2 Active N/A Not sooner than November 9, 2027 claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-mythos-5-1 Active N/A
The model-status table now lists claude-fable-5-2 and claude-mythos-5-2 as Active, with a tentative retirement date of not sooner than November 9, 2027, listed above the existing claude-fable-5-1 and claude-mythos-5-1 entries, which retain their September 1, 2027 date. Elsewhere the page renamed its two Amazon Bedrock model-table references from "Amazon Bedrock" and "Amazon Bedrock (Opus 4.6 and earlier)" to "Amazon Bedrock (bedrock-mantle)" and "Amazon Bedrock (bedrock-runtime)." 18 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-11-07.
2026-11-07 · roblox — Roblox rewrites AI Features responsibility and ownership language in its Terms of Use
Roblox Terms of Use (contains AI Features section) · detected weekly pass, 2026-11-07
Before:
”). This Virtual Content can be Traded for other Virtual Content or for a combination of Robux and Virtual Content . When Users engage in a Trade on the Services that includes the exchange of Robux , Roblox is entitled to a fee. However, if no Robux are exchanged as part of the Trade , Roblox will not receive any fee in connection with the transaction. (See here for a more detailed explanation of Roblox’s Trading System and the associated fees.) Removal of Virtual Content . Roblox has the right, in its sole discretion, to suspend the availability of, or remove from the Services and your Account , any content (including without limitation Experiences, Virtual Content , UGC Subscriptions , and UGC ) without advance notice. Roblox shall not be liable to any User for any losses you may experience because of such suspension or removal, and Roblox is not required to refund any Robux or other funds that a User has spent on any removed or suspended content, except where legally required. Paid Access. Paid Access experiences offered by a Creator require a one-time payment in Robux or local currency to access the experience (“Paid Access”). Paid Access experiences in Robux can only be purchased using Robux and Paid Access experiences in local currency can only be purchased using real currency. You can purchase a Paid Access experience on desktop in your local currency, or in USD if your local currency isn't available. You may request a refund for a Paid Access in local currency experie
After:
Community Standards explain how you can report Content that is illegal or violates the Roblox Terms including the Community Standards. The Content Moderation on Roblox
This pass found 65 changed spans in Roblox's Terms of Use. Among the listed spans, a renumbered "5.3.1 Responsibility for AI Features" clause now states Roblox is not responsible for AI Prompts and Outputs (the prior text stated the user, not Roblox, bears that responsibility), a new "5.3.2 Reliance on Outputs" section is added, a license clause now states Roblox does not assert ownership of intellectual property created in Outputs, and the document's appendix list of region-specific terms, the personal-information definition in the prohibited-conduct section, and the disclaimer text preceding the User Terms are each revised.
2026-11-07 · roblox — Roblox rewrites COPPA parental-rights language and several child-privacy provisions
Roblox Privacy and Cookie Policy · detected weekly pass, 2026-11-07
Before:
To make a request, please fill out our Customer Support Form . In the US and want to know more? We comply with the Children’s Online Privacy Protection Act (“COPPA”). To learn more about COPPA, you can visit the US Federal Trade
After:
To make a request, please fill out our Customer Support Form . We comply with the Children’s Online Privacy Protection Act (“COPPA”). COPPA affords parents the right to: Access your child’s information and correct it Request that Roblox delete your child’s information (If you ask us to do this, we may need to close their account because some information is necessary for Roblox’s service) Request that Roblox stop using your child’s information and withdraw any consent you previously granted When providing consent, receive direct notice of Roblox’s data notices. To exercise these rights, Roblox may require you to take action or provide additional information so we can verify your request, such as linking your account to your child’s account. To learn more about COPPA, you can visit the US Federal Trade
This pass found 53 changed spans in Roblox's Privacy and Cookie Policy; the excerpt above shows a new paragraph listing COPPA parental rights (access, deletion, stopping use, and direct notice) that was not previously present. Other listed spans change the parental email notification window from 24 hours to 7 days, reword its age-assurance method from a "selfie" used for "age assurance" to a "live selfie" used for "age checks," and remove an "under the age of 13" qualifier that had limited a persistent-identifier data-use restriction to users of that age.
2026-11-07 · perplexity — Perplexity's Terms of Service page drops its Privacy Notice update banner
Perplexity Terms of Service · detected weekly pass, 2026-11-07
Before:
of Service Products Business Customers Resources Pricing See a demo Try Perplexity We recently updated our consumer Privacy Notice. We added more detail about cookies, first-party advertising measurement, and your privacy choices. We also clarified that we do not sell your personal data or send your queries, prompts, or conversation content to advertisers. Perplexity Terms of Service Legal overview Platform User Terms Comet Terms of
After:
of Service Products Business Customers Resources Pricing See a demo Try Perplexity Perplexity Terms of Service Legal overview Platform User Terms Comet Terms of
A banner previously shown near the top of the Terms of Service page, announcing an update to Perplexity's consumer Privacy Notice and describing new detail on cookies, advertising measurement, and a data-sale clarification, is no longer present. The page's top navigation also added "Teammate" to its product list and "E-commerce" to its industry list.
2026-11-07 · openai — OpenAI adds gpt-5.5-cyber-preview deprecation notice
OpenAI API deprecations · detected weekly pass, 2026-11-07
Before:
deprecations are listed below, with the most recent announcements at the top. 2026-10-14: GPT-4 series model deprecations On October 14, 2026, we notified developers
After:
deprecations are listed below, with the most recent announcements at the top. 2026-11-05: GPT-5.5-Cyber The gpt-5.5-cyber-preview model is deprecated and will be removed from the API on February 1, 2027. Migrate to the most capable cyber model available to you before the shutdown date. Shutdown date Model / system Recommended replacement February 1, 2027 gpt-5.5-cyber-preview The most capable cyber model available to you. 2026-10-14: GPT-4 series model deprecations On October 14, 2026, we notified developers
A new entry dated 2026-11-05 announces that gpt-5.5-cyber-preview is deprecated and will be removed from the API on February 1, 2027, with a recommended replacement listed. The page also adds a sentence stating fine-tuned model versions follow their base model's shutdown schedule, and updates the gpt-image migration table's recommended replacements from gpt-image-2.5-sunburst/flare to gpt-image-2.6-sunburst/flare.
2026-11-07 · canva — Canva's policy archive index adds two new documents and several new dated entries
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-11-07
Before:
Current Recruitment Finder's Fee Addendum Current SMS Terms Current Jul 14, 2026 Show us how you use Canva Current Statement of Technical and Organisational
After:
Current Recruitment Finder's Fee Addendum Current SMS Terms Current Jul 14, 2026 Schedule 1 to the Canva API and App Developer Terms - Canva API Services Current Show us how you use Canva Current Statement of Technical and Organisational
The archive index added two documents not previously listed, "Canva's Electronic Signature Feature Terms" and "Schedule 1 to the Canva API and App Developer Terms - Canva API Services," each marked Current. It also added new dated prior-version entries for the Privacy Notice (Nov 2, 2026, listed twice), the Feature-Specific Terms (Oct 29, 2026), the Privacy Policy (Nov 2, 2026), and the Terms of Use (Oct 29, 2026).
2026-11-07 · apple — Apple adds iPad mini with A20 Pro to its Apple Intelligence device list
How to get Apple Intelligence (availability and capability) · detected weekly pass, 2026-11-07
Before:
and service is available. Refer to OpenAI for ChatGPT availability. Published Date: October 22, 2026
After:
and service is available. Refer to OpenAI for ChatGPT availability. Published Date: November 05, 2026
The device-compatibility list for iPad mini now reads "iPad mini (A17 Pro and A20 Pro)," adding the A20 Pro chip alongside the previously listed A17 Pro. The page's Published Date also moved from October 22, 2026 to November 05, 2026.
2026-11-07 · anthropic — Anthropic reorders the two partner platforms named in its deprecations page
Anthropic model deprecations · detected weekly pass, 2026-11-07
Before:
on Google Cloud (Anthropic managed) , and Microsoft Foundry . Partner-operated platforms, Amazon Bedrock and Claude Models on Google Cloud (Google managed), set their own retirement schedules, so a model’s lifecycle status and dates
After:
on Google Cloud (Anthropic managed) , and Microsoft Foundry . Partner-operated platforms, Claude Models on Google Cloud (Google managed) and Amazon Bedrock, set their own retirement schedules, so a model’s lifecycle status and dates
The sentence naming the partner-operated platforms that set their own retirement schedules now lists "Claude Models on Google Cloud (Google managed) and Amazon Bedrock," reversing the prior order ("Amazon Bedrock and Claude Models on Google Cloud (Google managed)"). This was the only changed span in the main content.
2026-10-31 · xai — xAI renames SpaceXAI LLC to SpaceX Super Intelligence LLC in its Terms of Service
xAI Terms of Service · detected weekly pass, 2026-10-31
Before:
the “ Service ”). These Terms form an agreement between you and SpaceXAI LLC (“ SpaceXSI ,” “ we ,” “ our ,” or “
After:
the “ Service ”). These Terms form an agreement between you and SpaceX Super Intelligence LLC (“ SpaceXSI ,” “ we ,” “ our ,” or “
The entity name "SpaceXAI LLC" was changed to "SpaceX Super Intelligence LLC" throughout the terms, including in the DSA representative and copyright lines. The "Last Updated" date changed from October 12 to October 27, 2026. 5 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-24.
2026-10-31 · xai — xAI renames SpaceXAI LLC to SpaceX Super Intelligence LLC in its Privacy Policy
xAI Privacy Policy · detected weekly pass, 2026-10-31
Before:
Grok Build Try for free English SpaceXSI Privacy Policy Last Updated: October 11, 2026 ( previous version ) At SpaceXAI LLC (" SpaceXSI ", " our ", " us " or "
After:
Grok Build Try for free English SpaceXSI Privacy Policy Last Updated: October 27, 2026 ( previous version ) At SpaceX Super Intelligence LLC (" SpaceXSI ", " our ", " us " or "
The entity name changed from "SpaceXAI LLC" to "SpaceX Super Intelligence LLC," including in the EU/UK contact section where "x.AI LLC" became "SpaceX Super Intelligence LLC." The "Last Updated" date changed from October 11 to October 27, 2026. 3 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-24.
2026-10-31 · replika — Replika rewrites its Terms of Service, including its domain, medical-disclaimer, and cancellation sections
Replika Terms of Service · detected weekly pass, 2026-10-31
Before:
resident, you have the right to withdraw from the agreement for the purchase of digital content without charge and without giving any reason within fourteen (14) days from the date of such agreement conclusion. The withdrawal right does not apply if the performance of the agreement has begun with your prior express consent and your acknowledgement that you thereby lose your right of withdrawal. You hereby expressly consent to the immediate performance of the agreement and acknowledge that you will lose your right of withdrawal from the agreement once our servers validate your purchase and the applicable purchase is successfully delivered to you. Therefore, you will not be eligible for a refund unless the digital content is defective. For users residing in Brazil: applicable law may allow you to withdraw from a contract within seven days from its signing or from the receipt of the service. If your transaction qualifies for a Legal Refund under the applicable law, we will process such refunds upon your request. # # # # # # 2.5. Device requirements To enjoy Replika via your smartphone or other Device, your
After:
you cancel before the trial ends, the subscription begins automatically at the trial’s end. 2.9. Device requirements To enjoy Replika via your smartphone or other Device, your
The terms changed across at least 58 spans, including a website and contact-email domain change from replika.ai to replika.com, a rewritten section 1 ("Our Services"/medical disclaimer) dropping its research-evidence sentence about Replika's conversation techniques, a registration section now referencing date-of-birth age verification, and replacement of the "Cancellation of Services" section with a new "Virtual Items and Consumables" section. The document's main region could not be extracted, so this capture is labeled main content unavailable rather than compared by region. 58 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-24.
2026-10-31 · replika — Replika rewrites its Privacy Policy, replacing face-and-head-movement data language with new location and biometric age-verification sections
Replika Privacy Policy · detected weekly pass, 2026-10-31
Before:
confidentiality measures, so to exercise your rights or make a request, please contact us as provided in Section 11.A below. 8. Use of Replika
After:
those third parties’ direct marketing purposes. To send such a request, please Contact Us. 9. Other US States Notice of Collection . To learn more about the categories of personal data we collect about you and how we use it, including how we disclose it to third parties and certain information about those third parties, please see Sections 1-3 above. Notice of Sale and Targeted Advertising . We “sell” your personal data and engage in targeted advertising by providing certain of your personal data to certain third parties, including through certain tracking technologies made available through our Services. Rights . You have a right to submit a request to access, know, portability, delete, correct, and to opt-out of the sale or processing of your personal data for targeted advertising purposes. You may exercise these rights in Section 7. If you live in certain states (e.g., Delaware, Oregon, or Minnesota), you have a right to request a list of specific third parties to which we’ve disclosed your personal data. If you live in Connecticut, you may have a right to request a list of specific third parties to which we have sold your personal data. And if you live in Maryland, you have a right to request a list of categories of third parties to which we’ve disclosed your personal data. 10. Additional information for your region European Economic Area and United Kingdom. The legal bases in Section 2.A apply to you. You may lodge a complaint with your local supervisory authority, or co
The policy changed across at least 39 spans, including replacing its description of face-and-head-movement/TrueDepth camera data collection with new "Location information" and "Biometric age-verification data" sections, and adding "Onboarding questions" as a new data category used to configure a user's Replika. It also changed its "Last updated" date from May 27 to October 28, 2026, and added a statement that advertising partners never access users' conversations or photos. 39 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-24.
2026-10-31 · perplexity — Perplexity Privacy Policy reformats its VeraSafe EU/UK contact details into prose
Perplexity Privacy Policy · detected weekly pass, 2026-10-31
Before:
under Article 27 of the GDPR, VeraSafe’s contact information is as follows: For EU inquiries: VeraSafe Ireland Ltd. Unit 3D North Point House North Point Business Park New Mallow Road, Cork T23AT2P Ireland +420 228 881 031 For UK inquiries: VeraSafe United Kingdom Ltd. 37 Albert Embankment London SE1 7TL United Kingdom +44 (20) 4532 2003 VeraSafe is also our data protection officer (DPO), in which capacity it may be contacted using the following details: VeraSafe, LLC c/o VeraSafe Ireland Ltd. Unit 3D North Point House North Point Business Park New Mallow Road, Cork T23AT2P Ireland +420 228 881 031 experts@verasafe.com If you are not satisfied with our response and are in the
After:
under Article 27 of the GDPR, VeraSafe’s contact information is as follows: for EU inquiries, VeraSafe Ireland Ltd., Unit 3D North Point House, North Point Business Park, New Mallow Road, Cork, T23AT2P, Ireland, +420 228 881 031; for UK inquiries, VeraSafe United Kingdom Ltd., 37 Albert Embankment, London SE1 7TL, United Kingdom, +44 (20) 4532 2003. VeraSafe is also our data protection officer (DPO), in which capacity it may be contacted using the following details: VeraSafe, LLC, c/o VeraSafe Ireland Ltd., Unit 3D North Point House, North Point Business Park, New Mallow Road, Cork, T23AT2P, Ireland, +420 228 881 031, experts@verasafe.com. If you are not satisfied with our response and are in the
The VeraSafe EU and UK Article 27 GDPR contact details, previously listed with line breaks between elements, are now written as comma-separated prose within sentences. The addresses, phone numbers, and DPO contact details themselves are unchanged. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-31 · openai — OpenAI model-deprecations page drops a fine-tuned-model shutdown table and confirms an older-model shutdown date
OpenAI API deprecations · detected weekly pass, 2026-10-31
Before:
October 23, 2026 ft-o4-mini-2025-04-16 gpt-5.6-terra October 23, 2026 o4-mini-2025-04-16 | o4-mini gpt-5.6-terra The following fine-tuned versions were also shut down on October 23, 2026: Shutdown date Model snapshot Recommended replacement base model October 23, 2026 ft-gpt-3.5-turbo gpt-5.6-terra October 23, 2026 ft-gpt-4 gpt-5.6-sol October 23, 2026 ft-gpt-4.1-nano-2025-04-14 gpt-5.6-luna October 23, 2026 ft-babbage-002 gpt-5.6-terra October 23, 2026 ft-davinci-002 gpt-5.6-terra 2026-04-22: Legacy GPT model snapshots (July 2026 shutdown) On April 22, 2026,
After:
October 23, 2026 ft-o4-mini-2025-04-16 gpt-5.6-terra October 23, 2026 o4-mini-2025-04-16 | o4-mini gpt-5.6-terra Fine-tuned versions of these models follow the same shutdown schedule as their underlying base models. 2026-04-22: Legacy GPT model snapshots (July 2026 shutdown) On April 22, 2026,
The page replaced its table of individual fine-tuned-model shutdown dates with a single sentence stating fine-tuned models follow their base model's shutdown schedule. It also changed the older-model deprecation note from a future "will be shut down" to a completed shutdown on September 28, 2026. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-31 · google — Gemini Apps Privacy Hub adds an avatar-sharing FAQ and Memory summary references
Gemini Apps Privacy Hub · detected weekly pass, 2026-10-31
Before:
channel (if you have one). Deleting one does not affect the other. Canvas What happens to my data when I interact with Canvas web
After:
channel (if you have one). Deleting one does not affect the other. What happens to my data when I share my avatar with other users? Be cautious and only share your avatar with people you trust. Once you share your avatar with another Google user, they will be able to create AI content featuring your likeness and/or voice, and you will not have control over what they generate using your avatar. You can revoke permission at any time. When you do, the user you stop sharing your avatar with will lose the ability to make more AI content using your avatar. But any content they have already created will not be deleted. When you share your avatar, the recipient will not have access to your original likeness recording and will not be able to edit or delete it. Canvas What happens to my data when I interact with Canvas web
The "Last updated" date changed from October 23 to October 26, 2026, and the page added a new FAQ on what happens to your data when you share your avatar with other users (including on revocation and recipient access). It also added references to a "Memory summary" feature and expanded the list of devices and places where Gemini is available, including Google TV and Navigation. 10 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-31 · canva — Canva Terms of Use renames its AI-products section and adds a Feature Specific Terms cross-reference
Canva Terms of Use · detected weekly pass, 2026-10-31
Before:
certain responsibilities you have in relation to your domain. 6. Use of AI Products You may use AI-powered features and functionality Canva may make available on
After:
certain responsibilities you have in relation to your domain. 6. Use of Products Your use of certain Canva products and features may be subject to additional obligations and restrictions in the Feature Specific Terms (opens in a new tab or window) . You may use AI-powered features and functionality Canva may make available on
The terms' effective date changed from 19 August 2026 to 30 October 2026. Section 6, previously titled "Use of AI Products," is now titled "Use of Products" and adds that use of certain Canva products may be subject to additional obligations in the Feature Specific Terms; the "Other policies" footer list also gained a link to "Canva's Electronic Signature Feature Terms." 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-31 · anthropic — Anthropic model-deprecations page makes punctuation corrections
Anthropic model deprecations · detected weekly pass, 2026-10-31
Before:
, top_p , top_k Removed in API version 2026-10-21 Returns a 400 error . Earlier API versions return it when set to a non-default value on
After:
, top_p , top_k Removed in API version 2026-10-21 Returns a 400 error. Earlier API versions return it when set to a non-default value on
Two straight apostrophes ("model's" and "60 days' notice") were changed to curly apostrophes, and an extra space before a period in "Returns a 400 error ." was removed. 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-31 · amazon — Amazon Alexa Terms of Use adds the Alexa Home Map App as a defined access point
Alexa Terms of Use · detected weekly pass, 2026-10-31
Before:
product or application that enables access to Alexa, such as Amazon Echo devices and the Alexa App. “Alexa Interactions” means all information related to your use of Alexa and
After:
product or application that enables access to Alexa, such as Amazon Echo devices, the Alexa App, and the Alexa Home Map App. "Alexa Home Map App” means the app provided by Amazon that provides access to the Alexa Home Map. “Alexa Interactions” means all information related to your use of Alexa and
The terms added the "Alexa Home Map App" to the list of products that enable access to Alexa and added a matching defined term for it. The "Last Updated" date changed from May 14 to October 28, 2026, and a typo ("number days") was corrected to "number of days." 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-31 · amazon — Amazon's Alexa privacy page describes a new Alexa.com chat-window voice transcription feature
Alexa, Echo Devices, and Your Privacy · detected weekly pass, 2026-10-31
Before:
how to transfer your Amazon account to another marketplace, go here . 6. Can I choose to automatically delete my voice recordings, text transcripts,
After:
how to transfer your Amazon account to another marketplace, go here . When you tap the microphone icon in the Alexa.com chat window and begin speaking, your speech is transcribed into text and displayed in the chat window. You can then review the transcription before submitting your request to Alexa, as if you had typed it. Once submitted, you can manage your request just like any typed request to Alexa. Your audio is transcribed in the cloud and then automatically deleted. 6. Can I choose to automatically delete my voice recordings, text transcripts,
The page added a description of tapping the microphone icon in the Alexa.com chat window to transcribe speech to text for review before submitting it to Alexa, stating audio is transcribed in the cloud and then automatically deleted. It also reworded two FAQ cross-reference titles, one from "How does Follow up Mode work?" to "How can I speak to Alexa without having to use or repeat the wake word?" and another describing how Echo devices support presence detection. 4 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-24.
2026-10-24 · perplexity — Perplexity rewrites its Privacy Notice's data rights and cross-border transfer sections
Perplexity Privacy Policy · detected weekly pass, 2026-10-24
Before:
categories of information transferred, by contacting us using the contact information below. We comply with the EU-U.S. Data Privacy Framework and the UK Extension to the EU-U.S. DPF as set forth by the U.S. Department of Commerce (collectively, the “DPF”) and have certified to the U.S. Department of Commerce that we adhere to the DPF Principles with regard to the processing of personal data received from the European Union and UK (and Gibraltar) in reliance on the DPF. If there is any conflict between the terms in this privacy policy and the DPF Principles, the Principles shall govern to the extent applicable to the information at issue. The Federal Trade Commission has jurisdiction over our compliance with the DPF and, in accordance with the DPF, we are responsible for onward transfers to third parties that process personal data subject to the DPF in a way that does not follow the DPF Principles. To learn more about the Data Privacy Framework program, and to view our certification, please visit the Data Privacy Framework website . We commit to resolve DPF Principles-related complaints about our collection and use of personal data. Individuals in the EU and UK with inquiries or complaints regarding our compliance with the DPF should first contact us, at support@perplexity.ai . If you have an unresolved complaint concerning our handling of personal data received in reliance on the DPF that we have not addressed satisfactorily, please contact our U.S.-based third-party dispute
After:
categories of information transferred, by contacting us using the contact information below. Third-Party Applications/Websites We may provide links to websites and other third-party content
The Privacy Policy removes its EU-U.S. Data Privacy Framework certification and compliance language and adds a new "Third-Party Applications/Websites" section, along with an expanded "Data Subject Rights" section describing self-service export and deletion and new contact channels. The policy's Last updated date changed from July 8, 2026 to October 23, 2026. 6 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · openai — OpenAI's Usage Policies add restrictions on using outputs to train competing models
OpenAI Usage Policies · detected weekly pass, 2026-10-24
Before:
medical essential government services product safety components national security migration law enforcement Changelog 2025-10-29: We’ve updated our Usage Policies to reflect a universal set
After:
essential government services product safety components national security, intelligence, and law enforcement migration Training and developing models. Capable models must have safeguards proportionate to the risks they pose before they are deployed. Training another model on OpenAI outputs can transfer capabilities without carrying those safeguards over. You must not use OpenAI models or their outputs to generate, prepare, select, or supply training material or signals for models that are intended as alternatives to or substitutes for OpenAI models or services. This includes: Creating or preparing training data, training tasks, rubrics, or reinforcement-learning environments. Providing feedback used in training, including labels, scores, rankings, preferences, rewards, or corrections. Selecting, filtering, or validating training data, tasks, or trajectories. You must not attempt to extract protected reasoning through unauthorized access, reverse engineering, or circumvention of safeguards. These restrictions apply whether you perform the work yourself, for another party, or through someone acting on your behalf, even if you do not train the resulting model. They also apply when outputs from product use or evaluation are later used or supplied for those training purposes. You may use OpenAI models for research, software development, model evaluation and benchmarking, and product development, within these boundaries. These restrictions do not prohibit uses expressly permitted b
A new section prohibits using OpenAI models or their outputs to create training data, rubrics, or reinforcement-learning environments, or to provide feedback, for models intended as alternatives to OpenAI's own, and prohibits extracting protected reasoning through reverse engineering or circumventing safeguards. The policy's effective date changed from October 29, 2025 to November 20, 2026. 8 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · openai — OpenAI moves the October 23, 2026 legacy GPT snapshot shutdown into Past deprecations
OpenAI API deprecations · detected weekly pass, 2026-10-24
Before:
models will be disabled only when the underlying base model is deprecated. 2026-04-22: Legacy GPT model snapshots To improve reliability and make it easier for developers to choose the right models, we are deprecating a set of older OpenAI models. Access to these models will be shut down on the dates below. Shutdown date Model snapshot Substitute model October 23, 2026 gpt-3.5-turbo-0125 | gpt-3.5-turbo , gpt-3.5-turbo-completions gpt-5.6-terra October 23, 2026 gpt-4-0613 | gpt-4 , gpt-4-0613-completions , gpt-4-completions gpt-5.6-sol October 23, 2026 gpt-4-1106-preview gpt-5.6-sol October 23, 2026 gpt-4-turbo | gpt-4-turbo-2024-04-09 , gpt-4-turbo-completions gpt-5.6-sol October 23, 2026 gpt-4.1-nano | gpt-4.1-nano-2025-04-14 gpt-5.6-luna October 23, 2026 gpt-4o-2024-05-13 gpt-5.6-sol October 23, 2026 gpt-image-1 gpt-image-2.5-sunburst or gpt-image-2.5-flare October 23, 2026 o1-2024-12-17 | o1 gpt-5.6-sol October 23, 2026 o1-pro-2025-03-19 | o1-pro gpt-5.6-sol ( reasoning.mode: pro ) October 23, 2026 o3-mini-2025-01-31 | o3-mini gpt-5.6-sol October 23, 2026 ft-o4-mini-2025-04-16 gpt-5.6-terra October 23, 2026 o4-mini-2025-04-16 | o4-mini gpt-5.6-terra We are also removing fine-tuned versions as below: Shutdown date Model snapshot Recommended replacement base model October 23, 2026 ft-gpt-3.5-turbo gpt-5.6-terra October 23, 2026 ft-gpt-4 gpt-5.6-sol October 23, 2026 ft-gpt-4.1-nano-2025-04-14 gpt-5.6-luna October 23, 2026 ft-babbage-002 gpt-5.6-terra October 23, 2026 ft-davi
After:
models will be disabled only when the underlying base model is deprecated. Past deprecations Past deprecations are listed below, with the most recent announcements
The legacy GPT model snapshots previously listed as shutting down October 23, 2026 are now listed under "Past deprecations" as having been shut down on that date, and explicit snapshot-to-replacement version mappings (e.g. gpt-5.4-nano, gpt-5-pro) were added throughout the deprecation tables. 10 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · google — Google's Gemini Privacy Hub adds a Memory summary FAQ entry
Gemini Apps Privacy Hub · detected weekly pass, 2026-10-24
Before:
remote computer data and settings . You can manage this setting anytime. How can I control whether Google uses my audio and Gemini Live
After:
remote computer data and settings . You can manage this setting anytime. What does the Memory summary setting control? Depending on eligibility and your settings, you can use memory summary to view, edit, and delete what Gemini Apps understands about you. Memory Summary helps make sure that Gemini Apps responses are accurate and relevant for you. You’re in control You control the contents of your memory summary. You can add new details, fix inaccuracies, or delete information you no longer want saved. You can also turn memory Summary on or off at any time. You can access your memory summary at any time in your settings . You can also view your memory summary with the prompt, "What do you know about me?". How can I control whether Google uses my audio and Gemini Live
A new FAQ entry explains the Memory summary setting, which lets users view, edit, or delete what Gemini Apps has stored about them and can be turned off, including by asking "What do you know about me?". The page's Last updated date changed from October 13, 2026 to October 23, 2026. 12 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · apple — Apple adds iPhone Duo to the Apple Intelligence device list and a storage note
How to get Apple Intelligence (availability and capability) · detected weekly pass, 2026-10-24
Before:
to free up storage on your iPhone, iPad , or Mac . Supported languages and regions With iOS 27, iPadOS 27, and macOS 27,
After:
to free up storage on your iPhone, iPad , or Mac . Storage used by Apple Intelligence may temporarily exceed storage requirements during and shortly after a software update, as new models are downloaded and older ones are removed. Supported languages and regions With iOS 27, iPadOS 27, and macOS 27,
The compatible-device list now includes iPhone Duo alongside existing iPhone, iPad, and Mac models, and a new sentence states that storage used by Apple Intelligence may temporarily exceed requirements during and after a software update as models are downloaded and removed. The page's Published Date changed from September 14, 2026 to October 22, 2026. 7 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · anthropic — Anthropic reorders high-risk use case sections in the Usage Policy
Anthropic Usage Policy · detected weekly pass, 2026-10-24
Before:
Pricing Log in to Claude Models Mythos Fable Opus Sonnet Haiku Solutions AI agents Code modernization Coding Commerce Customer support Cybersecurity Enterprise Financial advisors Financial services Government Healthcare Higher education K-12 teachers Legal Life sciences Nonprofits Sales Small business Claude Platform Overview Developer docs Pricing Ecosystem Marketplace Regional compliance Claude on AWS Google Cloud Microsoft Foundry Console login Resources Blog Claude partner network Community Connectors Courses Customer stories Developer blog Engineering at Anthropic Events Plugins Powered by Claude Service partners Tutorials Use cases Programs Campus Frontier Academy Scientists Startups Help and security Availability Status Support center Company Anthropic Careers Leadership Policy
After:
mental health safeguards on Claude.ai. Research Policy Commitments Company Products Models Solutions Help and security Terms and policies Our mission AI research and products
The Usage Policy's "High-risk Use Case Requirements" and "Additional Use Case Guidelines" sections were relocated relative to "Do Not Violate" in the document's structure; the page's header metadata and global footer navigation were also restyled. 5 changed spans on the page; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · anthropic — Anthropic adds a Claude Haiku 5.5 model report section to the Transparency page
Anthropic Transparency page · detected weekly pass, 2026-10-24
Before:
describe our safeguards further in our most recent Risk Report . Claude Opus 5.5 Summary Table Model description Claude Opus 5.5 is our new
After:
describe our safeguards further in our most recent Risk Report . Claude Haiku 5.5 Summary Table See Claude Haiku 5.5 System Card Model description Claude Haiku 5.5 is our fastest, most capable small model, built for high-volume work like summarization, subagents, and browser use Benchmarked Capabilities See our Claude Haiku 5.5 system card ’s Section 8 on capabilities. Acceptable Uses See our Usage Policy Release date October 2026 Modalities Claude Haiku 5.5 can understand both text (including voice dictation) and image inputs, engaging in conversation, analysis, coding, and creative tasks. Claude can output text, including text-based artifacts, and diagrams. Model architecture and training methodology Claude Haiku 5.5 was pretrained on large, diverse datasets to acquire language capabilities. After the pretraining process, Haiku 5.5 underwent substantial post-training, with the goal of making it an effective assistant whose behavior aligns with the values described in Claude’s constitution . Training Data Claude Haiku 5.5 was trained on a proprietary mix of publicly available information from the internet, public and private datasets, user data included in feedback or bugs or which users have explicitly permitted for training, and other sources, such as synthetic data generated by other models. In the course of its training, we used several data cleaning and filtering methods, including deduplication and classification. Testing Methods and Results Based on our assessments, w
A new Claude Haiku 5.5 summary table was added to the Model Report, covering its described capabilities, release date (October 2026), modalities, training methodology, training data sources, and testing methods. 37 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-24 · anthropic — Anthropic removes temperature, top_p, and top_k parameters in API version 2026-10-21
Anthropic model deprecations · detected weekly pass, 2026-10-24
Before:
Anthropic occasionally deprecates request parameters that no longer apply to current models. How the API treats a deprecated parameter depends on the model, as the following table shows. Most SDKs keep deprecated parameters in their request types so existing code continues to type-check. The Python SDK (v1.0 and later) removes temperature , top_p , and top_k , so passing them raises a TypeError . Parameter Status Behavior Recommended replacement temperature , top_p , top_k Deprecated (Claude 4.7 and later models) Returns a 400 error when set to a non-default value on Claude 4.7 and later models
After:
Anthropic occasionally deprecates request parameters that no longer apply to current models. A new model can reject a deprecated parameter, and a new API version can remove it for every model. Parameter Status Behavior Recommended replacement temperature , top_p , top_k Removed in API version 2026-10-21 Returns a 400 error . Earlier API versions return it when set to a non-default value on Claude 4.7 and later models
The temperature, top_p, and top_k parameters moved from deprecated to removed in API version 2026-10-21, where they now return a 400 error unconditionally; earlier API versions still return that error only when the parameters are set to a non-default value on Claude 4.7 and later models. 15 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-17.
2026-10-17 · xai — xAI Terms of Service: entity renamed from SpaceXAI to SpaceXSI
xAI Terms of Service · detected weekly pass, 2026-10-17
Before:
Grok Build Try for free Terms of Service - Consumer Last Updated: September 11, 2026 ( previous version ) Welcome to SpaceXAI These Terms of Service (“Terms”) apply to your or others’ access, interactions and/or use of Grok, Grokipedia, and SpaceXAI’s other services for individuals, including associated applications, features, tools, software and websites (collectively, the “Service”). These Terms form an agreement between you and SpaceXAI LLC (“SpaceXAI,” “we,” “our,” or “us”) when you accept these Terms or when you otherwise access, interact with, and/or use the platform. SpaceXAI LLC is a Nevada company, with its registered offices at 800 W Cesar Chavez St., Austin, TX 78701, USA. By accessing
After:
Grok Build Try for free Terms of Service - Consumer Last Updated: October 12, 2026 ( previous version ) Welcome to SpaceXSI These Terms of Service (“ Terms ”) apply to your or others’ access, interactions and/or use of Grok, Grokipedia, and SpaceXSI’s other services for individuals, including associated applications, features, tools, software and websites (collectively, the “ Service ”). These Terms form an agreement between you and SpaceXAI LLC (“ SpaceXSI ,” “ we ,” “ our ,” or “ us ”) when you accept these Terms or when you otherwise access, interact with, and/or use the platform. SpaceXAI LLC is a Nevada company, with its principal place of business at 800 W Cesar Chavez St., Austin, TX 78701, USA. By accessing
The company name used throughout the document changed from "SpaceXAI" to "SpaceXSI," including in the entity name, product references, and contact information, while the underlying entity remains "SpaceXAI LLC." The "Last Updated" date changed from September 11, 2026 to October 12, 2026, and "registered offices" was replaced with "principal place of business" in the entity's address clause. 52 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-10.
2026-10-17 · xai — xAI Privacy Policy: entity renamed to SpaceXSI; prompt guidance removed
xAI Privacy Policy · detected weekly pass, 2026-10-17
Before:
our mission, we have developed “Grok,” a conversational generative AI powered by SpaceXAI’s large language models. More information about SpaceXAI’s development and training of Grok and data controls is available in our Consumer FAQs , Enterprise FAQs , and SpaceXAI website . Please note that SpaceXAI is a separate company from X Corp. ("X", previously Twitter). 2. Personal information we collect (Notice at collection) We ask that you do NOT include personal information in your prompts and inputs into our Service; however, we cannot control what you provide to us. We may collect personal information from you and about you. Some examples
After:
our mission, we have developed “Grok,” a conversational generative AI powered by SpaceXSI’s large language models. More information about SpaceXSI’s development and training of Grok and data controls is available in our Consumer FAQs , Enterprise FAQs , and SpaceXSI website . Please note that SpaceXSI is a separate company from X Corp. ("X", previously Twitter). 2. Personal information we collect (Notice at collection) We may collect personal information from you and about you. Some examples
The company name changed from "SpaceXAI" to "SpaceXSI" throughout, and the date label changed from "Effective: August 24, 2026" to "Last Updated: October 11, 2026." Section 2's instruction asking users not to include personal information in prompts was removed, and new language was added elsewhere stating the company cannot control what users provide and that users are responsible for actions, including providing payment information, as if taken themselves. 20 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-10.
2026-10-17 · xai — xAI Acceptable Use Policy: entity renamed from SpaceXAI to SpaceXSI
xAI Acceptable Use Policy · detected weekly pass, 2026-10-17
Before:
X Grok Bot Developers API Console Documentation Grok Build Try for free SpaceXAI Acceptable Use Policy Effective: August 14, 2026 ( previous version ) SpaceXAI’s Acceptable Use Policy (“AUP”) applies to anyone using our Service, including consumers,
After:
X Grok Bot Developers API Console Documentation Grok Build Try for free SpaceXSI Acceptable Use Policy Effective: August 14, 2026 ( previous version ) SpaceXSI’s Acceptable Use Policy (“AUP”) applies to anyone using our Service, including consumers,
The company name used throughout the document changed from "SpaceXAI" to "SpaceXSI," in the policy name, body references, brand/social links, and contact references. The policy's effective date (August 14, 2026) is unchanged. 7 changed spans on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-10.
2026-10-17 · openai — OpenAI Service Terms add a Prompt-to-Voice provision
OpenAI Service Terms · detected weekly pass, 2026-10-17
Before:
End Users only, use APIs in accordance with the applicable documentation at https://platform.openai.com/docs (opens in a new window) . 2. Beta Services This section governs your use of services or features
After:
End Users only, use APIs in accordance with the applicable documentation at https://platform.openai.com/docs (opens in a new window) (b) Prompt-to-Voice. Prompt-to-Voice allows Customers to create synthetic voices from text prompts for use in generating speech. You may not use Prompt-to-Voice to reproduce any person’s voice without express consent and all necessary rights. Customer must clearly disclose that generated speech is AI-generated, must not intentionally remove or alter OpenAI-provided indicators of AI origin, and must require End Users to comply with these requirements. OpenAI’s indemnification obligations under these Service Terms and the Agreement do not apply to Prompt-to-Voice or its Output. 2. Beta Services This section governs your use of services or features
The Service Terms' effective date changed from September 29, 2026 to October 16, 2026. A new "(b) Prompt-to-Voice" subsection was added describing a synthetic-voice generation feature, consent and AI-disclosure requirements for its use, and a carve-out stating OpenAI's indemnification obligations do not apply to Prompt-to-Voice or its output. 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-10.
2026-10-17 · openai — OpenAI model deprecations page adds GPT-4 series shutdown notice
OpenAI API deprecations · detected weekly pass, 2026-10-17
Before:
models will be disabled only when the underlying base model is deprecated. Past deprecations Past deprecations are listed below, with the most recent announcements at the top. 2026-09-11: GPT-5.4-Cyber The gpt-5.4-cyber model is deprecated and will be removed from the API on October 1, 2026. Migrate to the most capable cyber model available to you before the shutdown date. Shutdown date Model / system Recommended replacement Oct 1, 2026 gpt-5.4-cyber The most capable cyber model available to you. 2026-05-08: gpt-5.2-chat-latest and gpt-5.3-chat-latest model snapshots On May 8th, 2026, we notified developers using gpt-5.2-chat-latest and gpt-5.3-chat-latest model snapshots of their deprecation and removal from the API. Shutdown date Model / system Recommended replacement Aug 10, 2026 gpt-5.2-chat-latest gpt-5.6-sol Aug 10, 2026 gpt-5.3-chat-latest gpt-5.6-sol 2026-04-22: Legacy GPT model snapshots To improve reliability and make it easier
After:
models will be disabled only when the underlying base model is deprecated. 2026-04-22: Legacy GPT model snapshots To improve reliability and make it easier
A new entry announces that on October 14, 2026, OpenAI notified developers of the deprecation of certain GPT-4 series models, with removal from the API on January 14, 2027. The page also added documentation for an optional shutdown_date field on the List models endpoint, and the 2026-09-11 GPT-5.4-Cyber deprecation entry moved to a different position in the list. 5 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-10.
2026-10-17 · google — Google Gemini Privacy Hub renames "Gemini Spark" to "Tasks in Gemini"
Gemini Apps Privacy Hub · detected weekly pass, 2026-10-17
Before:
Gemini Apps. For example, some Gemini responses are grounded on Search results. And when Gemini gets help from Google Assistant for certain quick voice actions, relevant Assistant settings apply. If you are 18 or older and upgrade to Gemini Apps, your
After:
Gemini Apps. For example, some Gemini responses are grounded on Search results. Gemini also relies on some Google settings, for example, Gemini can help you hands-free based on your “Hey Google” & Voice Match settings. If you are 18 or older and upgrade to Gemini Apps, your
The feature previously called "Gemini Spark" is now called "Tasks in Gemini" throughout the page, including in descriptions of how Gemini uses data with the remote browser and remote computer. The page's last-updated dates changed from September 24, 2026 and June 29, 2026 to October 13, 2026, and references to the Google settings governing Gemini's data use were updated, including replacing "Location History" with "Search Services History" and adding "Timeline" and "Hey Google & Voice Match" as relevant settings. 18 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-10.
2026-10-17 · canva — Canva Policy Archives index adds a new API and App Developer Terms version
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-10-17
Before:
AI Beta Program Terms Current Canva API and App Developer Terms Current Dec 23, 2025 Canva Apps Terms Current Canva Billboard Competition Guidelines and
After:
AI Beta Program Terms Current Canva API and App Developer Terms Current Oct 16, 2026 Dec 23, 2025 Canva Apps Terms Current Canva Billboard Competition Guidelines and
A new dated entry, October 16, 2026, was added to the archive listing for the Canva API and App Developer Terms, ahead of the existing December 23, 2025 entry. A second span shows an additional "Current" entry appearing near the APAC-region terms and Purchase Order Terms listings. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-10.
2026-10-10 · samsung — Samsung Galaxy AI processing page updates the minimum OS version and supported device list
Galaxy AI features: on-device versus cloud processing · detected weekly pass, 2026-10-10
Before:
Edge, S25+, S25, S25 FE, S24 Ultra, S24+, S24, S24 FE, Tab A11+, Tab S11 Ultra, Tab11, Tab S10+, Tab S10 Ultra, Tab S10
After:
Edge, S25+, S25, S25 FE, S24 Ultra, S24+, S24, S24 FE, Tab S12+, Tab S12 Ultra, Tab A11+, Tab S11 Ultra, Tab11, Tab S10+, Tab S10 Ultra, Tab S10
Two listed spans change the stated minimum software from "Android 14 with One UI 6.1" to "Android 15 with One UI 7.0"; the third adds "Tab S12+" and "Tab S12 Ultra" to the device list between the S24 FE phone and the Tab A11+ tablet. 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · openai — OpenAI API deprecations page removes the GPT-5.4-Cyber deprecation notice
OpenAI API deprecations · detected weekly pass, 2026-10-10
Before:
tts-1-hd gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-03-20 gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-12-15 gpt-realtime-2.1-mini 2026-09-11: GPT-5.4-Cyber The gpt-5.4-cyber model is deprecated and will be removed from the API on October 1, 2026. Migrate to the most capable cyber model available to you before the shutdown date. Shutdown date Model / system Recommended replacement Oct 1, 2026 gpt-5.4-cyber The most capable cyber model available to you. 2026-08-26: Transcription models On August 26, 2026, we notified developers using whisper-1
After:
tts-1-hd gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-03-20 gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-12-15 gpt-realtime-2.1-mini 2026-08-26: Transcription models On August 26, 2026, we notified developers using whisper-1
The entry "2026-09-11: GPT-5.4-Cyber — The gpt-5.4-cyber model is deprecated and will be removed from the API on October 1, 2026" and its shutdown-date table row are present in the before capture and absent from the after capture. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · meta — Meta Privacy Policy page's "updated" banner about agentic AI capabilities is gone
Meta Privacy Policy · detected weekly pass, 2026-10-10
Before:
how we process your information Other policies and articles Settings Privacy Policy Updated on Jul 23, 2026 We've updated our Privacy Policy with details about agentic AI capabilities and how we personalise your AI experiences. View details What is the Privacy Policy and what does it cover? Effective July
After:
how we process your information Other policies and articles Settings Privacy Policy What is the Privacy Policy and what does it cover? Effective July
The before capture carried a banner reading "Updated on Jul 23, 2026. We've updated our Privacy Policy with details about agentic AI capabilities and how we personalise your AI experiences. View details"; that banner is absent in the after capture, and no other wording differs in the listed span. 1 changed span on the page (no usable main region); the excerpt shows the largest. Main content: unavailable. Last unchanged 2026-10-03.
2026-10-10 · google — Gemini Apps Privacy Hub replaces its "saved info" FAQ with a cross-conversation memory description
Gemini Apps Privacy Hub · detected weekly pass, 2026-10-10
Before:
can turn that setting off in your Google Home History . How does Gemini Apps use my saved info and instructions? Gemini Apps use your instructions ("Saved info" in some regions) to customize
After:
can turn that setting off in your Google Home History . How do I manage what Gemini remembers about me and customize my responses? Depending on eligibility and your settings, Gemini can remember specific facts, interests, and preferences across conversations so you don’t have to repeat info in new chats. You can ask Gemini to remember details simply by talking to it. Gemini Apps use your instructions ("Saved info" in some regions) to customize
The FAQ question "How does Gemini Apps use my saved info and instructions?" is replaced with "How do I manage what Gemini remembers about me and customize my responses?", with new text stating that, depending on eligibility and settings, Gemini can remember facts, interests, and preferences across conversations. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · canva — Canva Policy Archives index assigns a date to the Billboard Competition Guidelines entry
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-10-10
Before:
Canva Apps Terms Current Canva Billboard Competition Guidelines and Terms & Conditions Current Canva Button Partner Agreement Current Canva Community Awards 2023 Current Canva Community
After:
Canva Apps Terms Current Canva Billboard Competition Guidelines and Terms & Conditions Jul 11, 2022 Canva Button Partner Agreement Current Canva Community Awards 2023 Current Canva Community
The index's "Canva Billboard Competition Guidelines and Terms & Conditions" row changes from "Current" to "Jul 11, 2022", while the adjacent "Canva Button Partner Agreement" and "Canva Community Awards 2023" rows remain listed as Current. 1 changed span in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · anthropic — Anthropic Usage Policy revises its effective date and multiple use-case provisions
Anthropic Usage Policy · detected weekly pass, 2026-10-10
Before:
related to using our products or services to automatically generate content and publish it for external consumption Additional Use Case Guidelines The below use cases
After:
or behavior, including through disciplining, suspending, or terminating a worker. Education and credentials — admitting, rejecting, or ranking applicants; grading an exam or assessment that determines admission, graduation, licensure, or certification (outside of multiple-choice testing); granting, denying, or revoking a professional license or certification; making a determination of academic misconduct or suspending or expelling a student. Healthcare access — setting triage or waitlist priority; determining eligibility for, denying, or discontinuing a treatment, service, or care program. Public benefits and services — determining eligibility for, or the amount of, government benefits, assistance, or essential services, including suspension and termination. Legal status and adjudication — deciding or recommending the outcome of a civil, administrative, or arbitral matter; determining immigration, asylum, or citizenship status. Prohibition Reminder: Use of our products or services to make or suggest decisions in law enforcement or criminal justice processes is prohibited under our Universal Usage Standards and is not a High-risk AI Recommendation. High-risk Physical Actions When connecting our models to hardware that can take physical actions, users should assess and understand 1) the range of possible outcomes if something goes wrong, 2) the nature and extent of potential impact to individuals, groups or populations, and 3) the degree of control the user maintains relative t
The policy's stated effective date changes from September 15, 2025 to November 12, 2026. The listed spans show revised wording throughout the Universal Usage Standards and Additional Use Case Guidelines sections, including the excerpted High-risk AI Recommendation categories (employment, education, healthcare, public benefits, and legal adjudication) and changes to the illegal-activity, unauthorized-access, weapons, discrimination, child-safety, and fraud provisions. 15 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · anthropic — Anthropic model-deprecations table adds claude-haiku-5-5 and renames a listed replacement model
Anthropic model deprecations · detected weekly pass, 2026-10-10
Before:
2026 June 15, 2026 claude-3-7-sonnet-20250219 Retired October 28, 2025 February 19, 2026 claude-haiku-4-5-20251001 Active N/A Not sooner than October 15, 2026 claude-3-5-haiku-20241022 Retired December
After:
2026 June 15, 2026 claude-3-7-sonnet-20250219 Retired October 28, 2025 February 19, 2026 claude-haiku-5-5 Active N/A Not sooner than October 7, 2027 claude-haiku-4-5-20251001 Active N/A Not sooner than October 15, 2026 claude-3-5-haiku-20241022 Retired December
The table adds a claude-haiku-5-5 row (Active, retirement not sooner than October 7, 2027) ahead of the existing claude-haiku-4-5-20251001 row, and five rows' listed replacement model changes from claude-sonnet-4-6 to claude-sonnet-5-5. The top_k deprecation note's stated applicability also changes from "Claude Opus 4.7 and later" to "Claude 4.7 and later models." 6 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-10 · amazon — Alexa privacy page adds a feedback-retention disclosure
Alexa, Echo Devices, and Your Privacy · detected weekly pass, 2026-10-10
Before:
Alexa Communication, and answer questions about attachments you’ve shared with Alexa. If your request was processed by an Alexa skill, deleting your voice recordings
After:
Alexa Communication, and answer questions about attachments you’ve shared with Alexa. If you report an issue or provide feedback to Alexa, for example by saying “Alexa, report an issue” or “Alexa, I have feedback,” or clicking on the thumbs down icon next to a chat in the Alexa app, we will retain a copy of text transcripts, typed requests, and other information related to your feedback to help us continue improving our services. If your request was processed by an Alexa skill, deleting your voice recordings
The new text states that reporting an issue or giving feedback to Alexa (by voice, or the thumbs-down icon in the app) causes a copy of text transcripts, typed requests, and related feedback information to be retained to improve services. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-10-03.
2026-10-03 · snap — Community Guidelines add "Plans" to the list of covered creative tools
Snap Community Guidelines · detected weekly pass, 2026-10-03
Before:
, Bitmojis, text, images, videos, generative AI, links or attachments, emojis, stickers, Lenses and other creative tools. It also applies to all behavior on Snapchat,
After:
, Bitmojis, text, images, videos, generative AI, links or attachments, emojis, stickers, Lenses, Plans, and other creative tools. It also applies to all behavior on Snapchat,
The scope list of tools the guidelines apply to now includes "Plans" alongside Lenses and other creative tools. The second changed span removes a stray space before a period in an unrelated sentence. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-09-26.
2026-10-03 · openai — Service Terms add sections on agentic features and Sign in with ChatGPT
OpenAI Service Terms · detected weekly pass, 2026-10-03
Before:
and personalized insights, is also subject to the Credit Score Terms .
After:
and personalized insights, is also subject to the Credit Score Terms . 14. Dots and Agentic Features (a) Responsibility. You decide how to use dots and related agentic features, including what access, authority, and instructions you give them. You are responsible for actions they take and content they share on your behalf. (b) Oversight. You are responsible for providing oversight appropriate to the task and its potential consequences. This includes setting clear instructions and appropriate limits and reviewing actions and communications. (c) Purchases and payments. You must have permission to use any account or payment method you provide. You are responsible for transactions made on your behalf. You must provide appropriate purchase guidance, monitor your accounts, and promptly raise any errors with the relevant financial institution or payment provider. Third-party terms may also apply to these transactions. 15. Sign in with ChatGPT Sign in with ChatGPT (“SIWC”) allows developers to enable users to connect their ChatGPT accounts, including certain features or functionality within those accounts, to their tool, application, or website. If you use SIWC in your services, the Sign in with ChatGPT Terms apply. If you are a user connecting your account to a service using SIWC, you agree to share information necessary to authenticate you (including name, email, and/or profile photo) and, where relevant, your ChatGPT usage. The information you share with another service will be su
The stated update date moved from September 21, 2026 to September 29, 2026, and new sections on agentic-feature ("dots") responsibility/oversight and on Sign in with ChatGPT account connections were added. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-09-26.
2026-10-03 · openai — API deprecations page adds two new notices and changes an image-model replacement
OpenAI API deprecations · detected weekly pass, 2026-10-03
Before:
are listed below, with the most recent announcements at the top. 2026-09-11: GPT-5.4-Cyber The gpt-5.4-cyber model is deprecated
After:
are listed below, with the most recent announcements at the top. 2026-10-01: GPT-5.3-Codex, GPT-5.1, GPT-5.4-Nano The following models are deprecated and will be removed from the API on April 1, 2027, with six months’ notice. Migrate to the recommended replacements before the shutdown date. Shutdown date Model / system Recommended replacement Apr 1, 2027 gpt-5.3-codex gpt-6-sol Apr 1, 2027 gpt-5.4-nano gpt-6-luna Apr 1, 2027 gpt-5.1 gpt-6-sol 2026-10-01: Text-to-speech models The following text-to-speech models are deprecated and will be removed from the API on January 6, 2027, with at least three months’ notice. Migrate to gpt-realtime-2.1-mini before the shutdown date. See the Realtime API guide to plan your migration. Shutdown date Model / system Recommended replacement Jan 6, 2027 tts-1 gpt-realtime-2.1-mini Jan 6, 2027 tts-1-hd gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-03-20 gpt-realtime-2.1-mini Jan 6, 2027 gpt-4o-mini-tts-2025-12-15 gpt-realtime-2.1-mini 2026-09-11: GPT-5.4-Cyber The gpt-5.4-cyber model is deprecated
Two deprecation notices dated 2026-10-01 were added: GPT-5.3-Codex, GPT-5.1 and GPT-5.4-Nano shut down April 1, 2027, and tts-1, tts-1-hd and two gpt-4o-mini-tts snapshots shut down January 6, 2027. The recommended replacement for gpt-image-1-mini, gpt-image-1.5, chatgpt-image-latest and gpt-image-1 changed from gpt-image-2 to "gpt-image-2.5-sunburst or gpt-image-2.5-flare." 7 changed spans in the main content; the excerpt shows the two new notices. Main content: changed. Last unchanged 2026-09-26.
2026-10-03 · microsoft — Services Agreement dated for 2026; Cortana section replaced with discontinuation notice
Microsoft Services Agreement · detected weekly pass, 2026-10-03
Before:
apply to your use of applicable Microsoft Services accessed through Cortana Services. Cortana provides information for your planning purposes only and you should exercise your own independent judgment when reviewing and relying on this information. Microsoft does not guarantee the reliability, availability or timeliness of personalized experiences provided by Cortana. Microsoft is not responsible if a Cortana feature delays or prevents you from receiving, reviewing or sending a communication or notification, or obtaining a service. iii. Third-Party Apps and Services. As part of delivering the Cortana Services, Cortana may suggest and help you interact with Third-Party Apps and Services (third-party skills or connected services). If you choose, Cortana may exchange information with Third-Party Apps and Services, such as your zip code and queries and responses returned by the Third-Party App and Services, to help you obtain requested services. Cortana may enable you to make purchases through Third-Party Apps and Services using the account preferences and settings you have established directly with those Third-Party Apps and Services. You can disconnect your Cortana Service from Third-Party Apps and Services at any time. Your use of Cortana Services to connect with Third-Party Apps and Services is subject to section 5 of these Terms. Publishers of Third-Party Apps and Services may change or discontinue the functionality or features of their Third-Party Apps and Services or integ
After:
and noncommercial use. As of 2024, Microsoft no longer supports Cortana Services. Microsoft 365 Apps and Services h. Microsoft 365 Apps and Services. i.
The stated Published and Effective dates moved from 2025 to 2026, and the Cortana Services description was replaced with "As of 2024, Microsoft no longer supports Cortana Services." Other changed spans include a section cross-reference number, product-list wording, and URL formatting; this excerpt shows the largest of 24 changed spans in the main content. Main content: changed. Last unchanged 2026-09-26.
2026-10-03 · anthropic — Transparency page adds Claude Sonnet 5.5 and Claude Opus 5.5
Anthropic Transparency page · detected weekly pass, 2026-10-03
Before:
Claude Opus 4 and Sonnet 4 Claude Opus 4.1 Claude Sonnet 3.7 Claude Fable 5.1 Summary Table Model description Claude Fable 5.1 is the
After:
Claude Opus 4 and Sonnet 4 Claude Opus 4.1 Claude Sonnet 3.7 Claude Sonnet 5.5 Summary Table Model description Claude Sonnet 5.5 is a faster, lower-cost complement to Claude Opus 5.5, and is strongest at well-scoped everyday tasks, fixing bugs, and creating polished documents, slides, and spreadsheets. Benchmarked Capabilities See our Claude Sonnet 5.5 system card ’s Section 8 on capabilities. Acceptable Uses See our Usage Policy Release date September 2026 Modalities Claude Sonnet 5.5 can understand both text (including voice dictation) and image inputs, engaging in conversation, analysis, coding, and creative tasks. Claude can output text, including text-based artifacts, and diagrams. Model architecture and training methodology Claude Sonnet 5.5 was pretrained on large, diverse datasets to acquire language capabilities. After the pretraining process, Sonnet 5.5 underwent substantial post-training, with the goal of making it an effective assistant whose behavior aligns with the values described in Claude’s constitution . Training Data Claude Sonnet 5.5 was trained on a proprietary mix of publicly available information from the internet, public and private datasets, and other sources, such as synthetic data generated by other models. The mix may also include user data from feedback or bug reports, or data that users have explicitly permitted for training. In the course of its training, we used several data cleaning and filtering methods, including deduplication and classifica
The page's Model Report date moved from August 17, 2026 to October 2, 2026, and Claude Sonnet 5.5 and Claude Opus 5.5 were added to the model list. A new summary-table entry describes Claude Sonnet 5.5, with a stated release date of September 2026. 3 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-09-26.
2026-10-03 · anthropic — Model deprecations page adds Claude Sonnet 4.5 retirement notice
Anthropic model deprecations · detected weekly pass, 2026-10-03
Before:
are listed in the following sections, with the most recent announcements first. 2026-06-05: Claude Opus 4.1 model This model was retired August
After:
are listed in the following sections, with the most recent announcements first. 2026-09-30: Claude Sonnet 4.5 model On September 30, 2026, Anthropic notified developers using Claude Sonnet 4.5 of its upcoming retirement on the Claude API. Retirement date Deprecated model Recommended replacement November 30, 2026 claude-sonnet-4-5-20250929 claude-sonnet-5-5 2026-06-05: Claude Opus 4.1 model This model was retired August
The active-models table added a claude-sonnet-5-5 row and changed claude-sonnet-4-5-20250929's status from Active to Deprecated. A new notice dated 2026-09-30 announces claude-sonnet-4-5-20250929's retirement from the API on November 30, 2026, with claude-sonnet-5-5 as the recommended replacement. 2 changed spans in the main content; the excerpt shows the largest. Main content: changed. Last unchanged 2026-09-26.
2026-09-26 · openai — OpenAI Service Terms add a sentence on credit score services
OpenAI Service Terms · detected weekly pass, 2026-09-26
Before:
Your use of Financial Services, including third-party data and content, is subject to the Financial Services Terms , including applicable data partner terms, disclaimers, and restrictions on use and redistribution.
After:
Your use of Financial Services, including third-party data and content, is subject to the Financial Services Terms , including applicable data partner terms, disclaimers, and restrictions on use and redistribution. Your use of credit score services, such as connecting to and accessing credit scores, reports, monitoring, and personalized insights, is also subject to the Credit Score Terms .
The "Updated" date changed from September 10, 2026 to September 21, 2026. One sentence was added after the Financial Services paragraph stating that use of credit score services is subject to the Credit Score Terms. Last unchanged 2026-09-19.
2026-09-26 · google — Gemini Apps Privacy Hub adds an agentic calling section
Gemini Apps Privacy Hub · detected weekly pass, 2026-09-26
Before:
Learn more about how you can use Gemini to get tasks done with screen automation . Shopping What happens to my data when I add items to the Google shopping cart from Gemini?
After:
Learn more about how you can use Gemini to get tasks done with screen automation . Agentic calling What happens to my data when I use Gemini's agentic calling capability? On eligible Pixel devices, the Gemini mobile app can make calls on your behalf to help with tasks, like booking reservations at a restaurant or checking product availability at a store. Learn how to use Gemini to make calls . You control the information Gemini will share on the call Before making a call, Gemini will confirm the information it plans to share with the other party. Depending on the task, this may include your name, email address, phone number, or other relevant details needed to complete your request. You must review this info and agree before Gemini will initiate the call. During a call, Gemini will not share sensitive personal information like credit card numbers, passwords, or health information. How data is used Google uses the transcripts from these calls to generate call summaries and to display your past calling activity within the Gemini app. We don’t use these audio recordings or transcripts to train Google's AI models. How data is stored On Google servers: Transcripts are sent to and securely stored on Google servers. Google does not store your audio recordings on its servers. On your device: A copy of the call transcript and audio recording is saved locally in your Phone by Google app's call history. You’re in control Your calling activity is available in your Gemini Apps Activity and used per the Gemini Apps Privacy Notice . You can find and delete it at any time. Learn how to delete your Gemini Apps activity . You can also delete your call history, transcripts, and recordings directly in the Phone by Google app. Shopping What happens to my data when I add items to the Google shopping cart from Gemini?
A new "Agentic calling" section was added, and the page's "Last updated" date changed from August 10, 2026 to September 24, 2026. In the Gemini Spark section, "connected folders" was added to the listed data sources in two places. Last unchanged 2026-09-19.
2026-09-26 · canva — Canva adds a Service Provider Data Processing Addendum to its policy archive index
Canva Policy Archives index (vendor-published prior versions) · detected weekly pass, 2026-09-26
Before:
Canva Service Provider Agreement Current Canva Sponsorship Addendum Current
After:
Canva Service Provider Agreement Current Sep 25, 2026 Canva Service Provider Data Processing Addendum Current Canva Sponsorship Addendum Current
A new row, "Canva Service Provider Data Processing Addendum," was added to the index with the date "Sep 25, 2026" and marked Current. A second date, "Sep 30, 2026," was added to the Feature-Specific Terms row ahead of the existing "Apr 14, 2026." Last unchanged 2026-09-19.
2026-09-26 · anthropic — Anthropic adds claude-opus-5-5 to the model status table and removes the Mythos Preview deprecation notice
Anthropic model deprecations · detected weekly pass, 2026-09-26
Before:
Model status Claude Mythos Preview ( claude-mythos-preview ) is deprecated. To migrate to Claude Mythos 5 ( claude-mythos-5 ), see the migration guide . Current and recently retired models are listed in the following table with their status: API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-mythos-5-1 Active N/A Not sooner than September 1, 2027 claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-mythos-5 Active N/A Not sooner than June 9, 2027 claude-mythos-preview Deprecated June 9, 2026 To be announced claude-opus-5 Active N/A Not sooner than July 24, 2027
After:
Model status Current and recently retired models are listed in the following table with their status: API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-mythos-5-1 Active N/A Not sooner than September 1, 2027 claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-mythos-5 Active N/A Not sooner than June 9, 2027 claude-mythos-preview Deprecated June 9, 2026 To be announced claude-opus-5-5 Active N/A Not sooner than September 22, 2027 claude-opus-5 Active N/A Not sooner than July 24, 2027
A row for claude-opus-5-5 was added to the model status table as Active, with a tentative retirement date of "Not sooner than September 22, 2027." The notice above the table stating that Claude Mythos Preview is deprecated, with a pointer to a migration guide for Claude Mythos 5, was removed; the claude-mythos-preview row is unchanged. Last unchanged 2026-09-19.
2026-09-19 · zoom — Zoom qualifies its do-not-sell statement and discloses business contact data sales
Privacy Statement · detected weekly pass, 2026-09-19
Before:
Opt out of the “sale” of personal data or “sharing” of personal data for targeted advertising purposes. We do not sell your personal data in the conventional sense. However, like many companies, we may use advertising and analytics services that are intended to analyze your interactions with our website or app
After:
Opt out of the “sale” of personal data or “sharing” of personal data for targeted advertising purposes. We do not sell your personal data governed by this Privacy Statement in the conventional sense. Sales of business contact related data may occur through our Zoom Common Room offering. Please see Zoom Common Room Privacy Statement for more information. However, like many companies, we may use advertising and analytics services that are intended to analyze your interactions with our website or app
The do-not-sell statement was qualified: "your personal data" became "your personal data governed by this Privacy Statement," and two sentences were added disclosing that sales of business contact related data may occur through Zoom Common Room, which is governed by a separate privacy statement. Zoom's own update notes date the change September 17, 2026, and the previous update note, dated July 27, 2026, was removed. Last unchanged 2026-09-12.
2026-09-19 · xai — xAI cuts its account inactivity termination window from over a year to 120 days
Terms of Service - Consumer · detected weekly pass, 2026-09-19
Before:
Your use of our Service could cause risk or harm to SpaceXAI, our users, or anyone else; or Your account has been inactive for over a year and you do not have a paid account. No Refund. Upon Service termination, you will not be entitled to any refund, except where required by law.
After:
Your use of our Service could cause risk or harm to SpaceXAI, our users, or anyone else; or Your account has been inactive for more than 120 days. No Refund. Upon Service termination, you will not be entitled to any refund, except where required by law.
The inactivity period after which xAI may terminate an account without notice fell from "over a year" to "more than 120 days," and the qualifier limiting that ground to accounts without a paid subscription was removed. The same section added "cancel your subscription or account" to the actions xAI may take at its sole discretion. No refund provision was changed. Last unchanged 2026-09-12.
2026-09-19 · snap — Snap adds AI-powered Team Snapchat replies to its generative AI feature list
Generative AI Disclosure · detected weekly pass, 2026-09-19
Before:
personalized titles or chapter labels for some Featured Stories in Memories , and generating songs from Chat messages . How can I recognize generative AI tools on Snapchat?
After:
personalized titles or chapter labels for some Featured Stories in Memories , generating songs from Chat messages , and AI-powered replies when you message Team Snapchat (where available). How can I recognize generative AI tools on Snapchat?
One feature was added to the list of Snapchat features that use generative AI: AI-powered replies when a user messages Team Snapchat, marked "(where available)." No other listed feature changed. Last unchanged 2026-09-12.
2026-09-19 · apple — Apple Intelligence privacy page adds a web search path and softens two data commitments
Apple Intelligence - Privacy · detected weekly pass, 2026-09-19
Before:
When your device sends a request to Private Cloud Compute, Apple only collects limited information about the request, such as the approximate size of the request and response, which features are used for the request, and how long the request takes to complete.
After:
When your request is processed on Private Cloud Compute or using Private Information Retrieval, Apple may collect limited information about the request, such as the approximate size of the request and response, which features are used for the request, and how long the request takes to complete.
The page was substantially rewritten, from 920 to 1,073 words across 34 changed spans. In the excerpt above, "Apple only collects limited information" became "Apple may collect limited information." Elsewhere, "When possible, Apple Intelligence models run entirely on device" became "In many cases"; a new passage describes Apple Intelligence using Search to retrieve up-to-date information from the web, enabled by a "Use Broad World Knowledge" setting in Shortcuts; a new passage describes Private Information Retrieval; and references to the "ChatGPT extension" were replaced with "Apple Intelligence Extensions." Last unchanged 2026-09-12.
2026-09-19 · apple — Apple adds daily usage limits to Apple Intelligence and says expanded access will carry a fee
Apple Intelligence - Availability · detected weekly pass, 2026-09-19
Before:
devices that support Apple Intelligence will be able to activate and use it in that region. Apple Intelligence features With iOS 26, iPadOS 26, and macOS Tahoe 26 and later, Apple Intelligence powers these features:
After:
devices that support Apple Intelligence will be able to activate and start using a core set of Apple Intelligence features in that region. New Apple Intelligence features will not be available in China while Apple works through regulatory requirements. Feature Usage Limits Certain Apple Intelligence features that rely on server-side models are subject to daily usage limits, including but not limited to Siri AI, Intelligent Photo Editing Tools, Image Playground, and AFM 3 Cloud models in Shortcuts. Daily limits may vary by feature, request complexity, system demand, system policies, and other factors. Expanded access to such features will be available for a fee in the future. Use of these features is subject to the Apple Intelligence Terms and Conditions. Learn more at Apple Intelligence Usage Terms . Feature availability by platform With iOS 27, iPadOS 27, macOS 27 Golden Gate, watchOS 27, and visionOS 27 the next generation of Apple Intelligence powers these features:
The page grew from 1,254 to 1,661 words across 98 changed spans. A new "Feature Usage Limits" section states that Apple Intelligence features relying on server-side models are subject to daily usage limits, naming Siri AI, Intelligent Photo Editing Tools, Image Playground, and AFM 3 Cloud models in Shortcuts, that limits may vary by feature, request complexity, system demand and system policies, and that "Expanded access to such features will be available for a fee in the future." A new sentence states that new Apple Intelligence features will not be available in China while Apple works through regulatory requirements, and that devices in a newly supported region can activate "a core set of" features rather than the prior unqualified wording. The feature matrix was rebuilt around iOS 27, iPadOS 27, macOS 27 Golden Gate, watchOS 27 and visionOS 27. Last unchanged 2026-09-12.
2026-09-19 · anthropic — Anthropic publishes Claude Fable 5.1 and Mythos 5.1 model cards, both deployed under ASL-3
Transparency & Annual Report · detected weekly pass, 2026-09-19
Before:
how we're addressing potential risks. Claude Opus 5 Claude Sonnet 5 Claude Fable 5 Claude Mythos 5 Claude Opus 4.8 Claude Opus 4.7
After:
how we're addressing potential risks. Claude Fable 5.1 Claude Mythos 5.1 Claude Opus 5 Claude Sonnet 5 Claude Fable 5 Claude Mythos 5 Claude Opus 4.8 Claude Opus 4.7
The page grew from 18,124 to 20,410 words. A 2,279-word insertion adds full model cards for Claude Fable 5.1 and Claude Mythos 5.1, stating a September 2026 release date, that both were deployed with ASL-3 protections and treated as having CB-1 capabilities, and that the two are the same underlying model differing only in safeguards, with Mythos 5.1 restricted to trusted access programs and Claude Security. The insertion reports RSP evaluation conclusions for chemical and biological weapons, novel weapons development, and automated AI research and development; safeguards evaluation results in the suicide and self-harm domain, including a stated pre-launch change to the claude.ai system prompt; computer-use attack success rates; and ExploitGym cyber and protein-design bio results. The alignment risk assessment is stated as moved "from very low to low" in the August 2026 Risk Report. Last unchanged 2026-09-12.
2026-09-19 · anthropic — Anthropic adds Claude Mythos models to the deprecation table, one already deprecated
Model Deprecations · detected weekly pass, 2026-09-19
Before:
API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-opus-5 Active N/A Not sooner than July 24, 2027
After:
API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-mythos-5-1 Active N/A Not sooner than September 1, 2027 claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-mythos-5 Active N/A Not sooner than June 9, 2027 claude-mythos-preview Deprecated June 9, 2026 To be announced claude-opus-5 Active N/A Not sooner than July 24, 2027
Three rows were added to the model status table: claude-mythos-5-1 and claude-mythos-5, both Active, with tentative retirement dates of September 1, 2027 and June 9, 2027; and claude-mythos-preview, listed as Deprecated on June 9, 2026 with a retirement date of "To be announced." No existing row changed. Last unchanged 2026-09-12.
2026-09-12 · snap — Snap adds content guidelines for algorithmic recommendation
Privacy Policy · detected weekly pass, 2026-09-12
Before:
self-expression, while striving to make sure Snapchatters can use our services safely every day. Creator Monetization Policy Content must adhere to the policies to be eligible for monetization. Advertising Policies These apply to all aspects of advertisements served by Snap.
After:
self-expression, while striving to make sure Snapchatters can use our services safely every day. Content Guidelines Additional standards for algorithmic recommendation beyond the creator's friends or subscribers Creator Monetization Policy Content must adhere to the policies to be eligible for monetization. Advertising Policies These apply to all aspects of advertisements served by Snap.
New "Content Guidelines" section added with language governing "algorithmic recommendation beyond the creator's friends or subscribers." Last unchanged between 2026-09-05 and 2026-09-12.
2026-09-12 · openai — Financial Services section added to Service Terms
Service Terms · detected weekly pass, 2026-09-12
Before:
12. OpenAI Academy OpenAI Academy is a free learning hub and community for users of OpenAI products and services. Use of OpenAI Academy is subject to OpenAI's Terms of Use and OpenAI Academy's Terms of Use and Code of Conduct. For users with accounts associated with a business or enterprise workspace, the OpenAI Services Agreement or other business or enterprise terms do not apply to use of OpenAI Academy, even if a user uses Sign in with ChatGPT with their workspace account. If you use an email address or account associated with an organization, including through Sign in with ChatGPT, OpenAI may share information about your OpenAI Academy activity, such as enrollment, engagement, progress, and completion status, with your organization's owners or administrators.
After:
12. OpenAI Academy OpenAI Academy is a free learning hub and community for users of OpenAI products and services. Use of OpenAI Academy is subject to OpenAI's Terms of Use and OpenAI Academy's Terms of Use and Code of Conduct. For users with accounts associated with a business or enterprise workspace, the OpenAI Services Agreement or other business or enterprise terms do not apply to use of OpenAI Academy, even if a user uses Sign in with ChatGPT with their workspace account. If you use an email address or account associated with an organization, including through Sign in with ChatGPT, OpenAI may share information about your OpenAI Academy activity, such as enrollment, engagement, progress, and completion status, with your organization's owners or administrators. 13. Financial Plugins and ChatGPT for Financial Services OpenAI's financial plugins and ChatGPT for Financial Services (together, "Financial Services") and their Output are for informational purposes only. OpenAI does not provide financial or investment advice through Financial Services. Data and Output may be inaccurate, delayed, or outdated. Review important information and exercise independent professional judgment before relying on it. Seek advice from a qualified financial advisor where appropriate. You remain responsible for decisions made using Financial Services. Your use of Financial Services, including third-party data and content, is subject to the Financial Services Terms, including applicable data partner terms, disclaimers, and restrictions on use and redistribution.
Effective date changed from June 12, 2026 to September 10, 2026. New section 13 governing financial plugins and Financial Services added, establishing limitations on financial advice and disclaimers. Last unchanged between 2026-09-05 and 2026-09-12.
2026-09-12 · openai — OpenAI adds GPT-5 model deprecation information
Model Deprecations · detected weekly pass, 2026-09-12
Before:
Deprecations | OpenAI API For the complete documentation index, see llms.txt. Models & API access Codex Overview Models Agents Tools Voice & Audio Production API reference Overview Models Agents Tools Voice & Audio Production API reference
After:
Deprecations | OpenAI API For the complete documentation index, see llms.txt. Models & API access Codex Overview Models Agents Tools Audio & voice Production API reference Overview Models Agents Tools Audio & voice Production API reference
Navigation label changed from "Voice & Audio" to "Audio & voice." Additional references to GPT-5 model deprecation information added (4 more instances of "gpt-5" and 3 more "2026" date references). Last unchanged between 2026-09-05 and 2026-09-12.
2026-09-12 · microsoft — Microsoft Copilot features expanded with Cowork Agents
Microsoft 365 Copilot for Individuals · detected weekly pass, 2026-09-12
Before:
Microsoft 365 Copilot Plans for Individuals | Microsoft 365 This is the Trace Id: b29a75aafa03cd0246cece9f38602bcf Skip to main content Microsoft 365 Copilot Work IQ AI chat Complete tasks with Cowork Agents for work Build your own agents Copilot in your apps Copilot vs the competition
After:
Microsoft 365 Copilot Plans for Individuals | Microsoft 365 This is the Trace Id: 12e73326f2010fbbcc628a1822a8996b Skip to main content Microsoft 365 Copilot Work IQ AI chat Complete tasks with Cowork Agents for work Build your own agents Copilot in your apps Copilot vs the competition Copilot vs ChatGPT Enterprise Copilot vs Gemini Enterprise Copilot vs Claude Enterprise Copilot for individuals
Trace ID updated (volatile token, normalized). Feature descriptions expanded with additional comparative information added. Content size increased by 4700 bytes, suggesting expanded feature documentation or capability descriptions. Last unchanged between 2026-09-05 and 2026-09-12.
2026-09-12 · anthropic — Anthropic adds Korea-specific privacy compliance disclosures
Privacy Policy · detected weekly pass, 2026-09-12
Before:
Privacy Policy Effective July 8, 2026 Previous Version English Anthropic is an AI safety and research company working to build reliable, interpretable, and steerable AI systems.
After:
Privacy Policy Effective September 10, 2026 Previous Version English Anthropic is an AI safety and research company working to build reliable, interpretable, and steerable AI systems. Supplemental Disclosures for Residents of the Republic of Korea This section applies to the extent that Anthropic is subject to the Personal Information Protection Act ("PIPA"), and supplements the information in the Privacy Policy. Items of Personal Data Collected and Purpose of Use: Name, birthdate, sex, native-foreigner-code, and DI (duplication information). Purpose: Identity verification, Fraud and abuse monitoring.
Effective date changed from July 8, 2026 to September 10, 2026. New section added governing PIPA compliance for Republic of Korea residents, including requirements for collecting identity verification data and managing cross-border personal data transfers. Last unchanged between 2026-08-28 and 2026-09-12.
2026-09-05 · xai — Forum-selection and class-action waiver clause added to the consumer terms
SpaceXAI Terms of Service - Consumer · detected weekly pass, 2026-09-05
Before:
If you reside in the European Economic Area, United Kingdom, or Switzerland (collectively, "Europe"), your use of the Service is governed in part by the Europe Specific Terms ("EST"). Please read our Privacy Policy , which describes how we handle personal information.
After:
If you reside in the European Economic Area, United Kingdom, or Switzerland (collectively, "Europe"), your use of the Service is governed in part by the Europe Specific Terms ("EST"). Unless prohibited by your country's applicable law, any disputes between you and us will be resolved by a court in Tarrant County or Wichita County, Texas, based on Texas law, except that if the forum selection clause in these Terms is unenforceable for any reason, then the dispute will be arbitrated. To the extent permitted by applicable law, you agree to waive any right to bring a class action or other representative action. Please read our Privacy Policy , which describes how we handle personal information.
The document's stated effective date moved from August 24, 2026 to September 1, 2026. The capture retained 2026-08-28 does not contain the dispute-resolution paragraph; the capture retained 2026-09-05 does. The intervening pass on 2026-08-31 confirmed the earlier text unchanged, so the change occurred between 2026-08-31 and 2026-09-05. The same revision added a registered-office line: "SpaceXAI LLC is a Nevada company, with its registered offices at 800 W Cesar Chavez St., Austin, TX 78701, USA." 139 changed spans in total; the remainder are section-numbering and capitalization edits.
2026-09-05 · anthropic — claude-fable-5-1 added to the model lifecycle table
Claude Model Deprecations · detected weekly pass, 2026-09-05
Before:
API model name Current state Deprecated Tentative retirement date claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-opus-5 Active N/A Not sooner than July 24, 2027
After:
API model name Current state Deprecated Tentative retirement date claude-fable-5-1 Active N/A Not sooner than September 1, 2027 claude-fable-5 Active N/A Not sooner than June 9, 2027 claude-opus-5 Active N/A Not sooner than July 24, 2027
One row added; no existing row's state or retirement date changed. Change interval 2026-08-31 to 2026-09-05. The pass counted 40 changed spans for this source, but the other 39 are navigation chrome and the disappearance of sixteen "Loading" placeholders present in the 2026-08-28 capture, not document text.
2026-08-31 · OpenAI — Transcription model deprecation announced for February 26, 2027
Model Deprecations · detected weekly pass, 2026-08-31
Before:
Upcoming deprecations are listed below, with the most recent announcements at the top. 2026-07-20: Legacy audio, realtime, and transcription models On July 20, 2026, we notified developers using legacy audio, realtime, and transcription model families and snapshots of their deprecation
After:
Upcoming deprecations are listed below, with the most recent announcements at the top. 2026-08-26: Transcription models On August 26, 2026, we notified developers using whisper-1 , gpt-4o-transcribe , gpt-4o-mini-transcribe , and gpt-4o-transcribe-diarize of their deprecation and removal from the API on February 26, 2027. For information about the recommended replacements, see the transcription guide . Shutdown date Model / system Recommended replacement Feb 26, 2027 whisper-1 gpt-live-transcribe or gpt-transcribe Feb 26, 2027 gpt-4o-transcribe gpt-live-transcribe or gpt-transcribe Feb 26, 2027 gpt-4o-mini-transcribe gpt-live-transcribe or gpt-transcribe Feb 26, 2027 gpt-4o-transcribe-diarize gpt-live-transcribe or gpt-transcribe 2026-07-20: Legacy audio, realtime, and transcription models
A new dated entry was added at the top of the upcoming-deprecations list, carrying a shutdown date of February 26, 2027 for four transcription models. In the same pair of captures the existing 2025-08-20 Assistants API entry was rewritten from "The Assistants API was officially sunset on August 26, 2026, following its deprecation announcement on August 26, 2025." to "On August 26th, 2025, we notified developers using the Assistants API of its deprecation and removal from the API one year later, on August 26, 2026."
2026-08-31 · Character.AI — Helpfulness vote counter changed
Safety Center · detected weekly pass, 2026-08-31
Before:
Was this article helpful? Yes No 462 out of 773 found this helpful Share via: Articles in this section: How do I block another user? EU Digital Services Act Online Safety for Australian Users Safety Center
After:
Was this article helpful? Yes No 465 out of 776 found this helpful Share via: Articles in this section: How do I block another user? EU Digital Services Act Online Safety for Australian Users Safety Center
The only difference between the two captures is the reader-feedback tally at the foot of the help-center article, which moved from 462 of 773 to 465 of 776. The stated update date, October 30, 2025, and the body of the Safety Center text are identical between the two captures.