Live feed · updated daily

Recent policy changes

2941 policy changes detected across 352 platforms. Most platforms don't announce policy changes — these updates were detected automatically.

Stay ahead of the changes

Start monitoring platform changes

Free: research letter. Monitor: same-day alerts on the platforms you choose.

352 Entities monitored
844 Documents tracked
2941 Changes detected
Showing all changes. Show significant changes only
September 4, 2026
Peacock
Peacock Privacy Policy
low
Formatting correction detected in privacy policy; no substantive policy changes.
Why it matters: This change has no operational significance. It is a typographical correction only. The substantive terms, data handling practices, and consumer protections outlined in the privacy policy remain unchanged.
Uniswap
Uniswap Privacy Policy
low
Privacy Policy metadata timestamp updated; no material policy changes detected.
Why it matters: This change has no operational significance for users or the company. It is a routine metadata refresh indicating the document was reviewed on or around the detected date.
Medium
Medium Terms of Service
low
Removed metadata from Terms of Service header; no substantive policy changes.
Why it matters: This change has no operational significance. The removal of numerical metadata from the document header does not alter any substantive terms, conditions, rights, or obligations that govern Medium's service.
OpenSea
OpenSea Privacy Policy
low
Cryptocurrency price display updated in privacy policy footer
Why it matters: This change has no operational significance for privacy rights, data handling, or user obligations. The update reflects a routine refresh of market data displayed in the policy document footer.
OpenSea
OpenSea Terms of Service
low
Price figure updated in Terms of Service display; no substantive policy changes detected.
Why it matters: This change does not materially affect the substantive terms or operational impact of OpenSea's Terms of Service. It is a numerical display update with no bearing on user rights, obligations, or platform governance.
Stay ahead of the changes

You're seeing a fraction of what's changing

ConductAtlas monitors every tracked platform and captures every policy update.

Coursera
Coursera Privacy Notice
low
Expanded data categories for personalization to explicitly include education, career, and skills information alongside identity, account, demographic, and device data.
Why it matters: The updated terms now explicitly state which data categories are used to personalize your learning experience, recommendations, and engagement tools. The addition of 'Education, Career and Skills Information' to the disclosed categories clarifies the scope of data Coursera processes for personalization purposes, enabling transparency about what information influences your experience on the platform.
Pinterest
Pinterest Terms of Service
low
Added table of contents to Terms of Service organizing 13 major policy sections for easier navigation.
Why it matters: The table of contents makes Pinterest's Terms of Service more navigable by providing a clear roadmap to major policy sections. This is a usability improvement that does not change the substantive terms users operate under.
Pinterest
Pinterest Privacy Policy
low
Restructured privacy policy with new table of contents and reorganized sections for transparency and navigation.
Why it matters: The restructured privacy policy improves transparency and accessibility by organizing disclosures into clearly labeled sections covering data collection, usage, legal bases, rights, sharing, transfers, retention, and children's information. Users can now more easily locate specific privacy practices and their rights under the policy.
OpenAI
OpenAI GPT-5.5 System Card
low
Restructured GPT-5.5 System Card navigation: removed links to Hugging Face incident and zero data retention docs, added GPT-6 Astra safety overview.
Why it matters: This change updates which safety and incident documentation is directly accessible from OpenAI's GPT-5.5 System Card. Users seeking guidance on specific prior security incidents or data retention policies will no longer find those links in this location; instead, the card directs readers toward newer model safety documentation. The shift reflects organizational prioritization of current model releases over archival of prior incident documentation.
OpenAI
OpenAI Frontier Governance Framework
low
Reorganized Frontier Governance Framework webpage; removed Hugging Face incident reference; updated featured content links.
Why it matters: OpenAI's Frontier Governance Framework serves as the company's primary vehicle for explaining how its safety and security practices align with emerging AI regulations such as California's Transparency in Frontier AI Act and the EU AI Act's Code of Practice. This update is a minor webpage reorganization that does not alter the substantive content or scope of the framework itself.
OpenAI
OpenAI Enterprise Privacy
low
Formatting change to Enterprise Privacy document; no material operational changes to data control or compliance commitments.
Why it matters: This is a formatting change with no operational significance. The substantive privacy commitments and data control frameworks remain identical to the prior version.
OpenAI
OpenAI EU Terms of Use
low
Removed language-selection header elements from Europe Terms of Use; no policy substance changed.
Why it matters: This change has no operational significance. It is a visual and structural update to how the terms document is presented, not a modification of the underlying policy commitments or user rights.
September 3, 2026
Google AI Studio
Gemini API Terms of Service
low
Updated model version reference from Gemini 3.7 Flash to Gemini 3.8 Flash in Gemini API Terms.
Why it matters: The updated terms clarify which Gemini model version is currently available through the API. For developers integrating the Gemini API, this ensures they reference the correct model version in their implementations.
Google Cloud
Google Cloud Terms
high
Expanded billing authority: can charge full-month fees for suspected bad-faith usage or payment avoidance, suspend services for delayed-payment customers deemed risky.
Why it matters: The updated terms grant Google Cloud substantially broader authority to impose billing charges and suspend or terminate services based on its subjective determination of customer behavior and financial risk, rather than requiring actual non-payment, documented breach, or opportunity to cure. This shift from objective triggering events to predictive determination creates material operational and financial exposure for customers, particularly those using delayed payment methods or running usage patterns Google might view as attempting to optimize costs or work around usage caps.
RunPod
RunPod Privacy Policy
low
Updated mailing address for data subject requests from New Jersey to San Francisco.
Why it matters: The updated address ensures that individuals submitting formal privacy requests under GDPR, CCPA, and similar regulations route their requests to the correct operational location. Providing accurate contact information is a basic requirement for enabling individuals to exercise privacy rights.
RunPod
RunPod Terms of Service
low
Updated contact and legal notice address from Moorestown, NJ to San Francisco, CA.
Why it matters: This change affects the procedural requirement for how and where customers must send formal legal notices or complaints to RunPod. Using the correct address ensures notices are properly received and processed according to the terms of service.
Hilton
Hilton Privacy Statement
low
Expands disclosed data categories that may be collected and shared with third parties to include gender, telephone, arrival time, departure date, and spending amount.
Why it matters: The updated language increases transparency about what guest data Hilton may collect and share with third parties, making explicit five specific data categories previously not listed in this section. This supports compliance with privacy regulations that require clear disclosure of what personal information organizations collect and with whom they share it, and allows guests to understand the scope of data that may flow to Hilton's partners.
Stay ahead of the changes

You're seeing a fraction of what's changing

ConductAtlas monitors every tracked platform and captures every policy update.

DeepL
DeepL Terms and Conditions
low
Removed 'Cancel subscriptions here' navigation link from Terms and Conditions footer.
Why it matters: The removal of a direct 'Cancel subscriptions here' link from the website footer may affect user discoverability of the subscription cancellation process. Under consumer protection law in many jurisdictions, cancellation mechanisms must be available as easily as the signup process; if this was the primary cancellation access point, the organization should ensure an equally conspicuous alternative is in place.
DeepL
DeepL Privacy Policy
low
Removed subscription cancellation reference from privacy policy.
Why it matters: The change is minor editorial maintenance. Removal of a single sentence about subscription procedures from a privacy policy does not materially alter user rights, data handling disclosures, or regulatory compliance posture.
Mistral AI
Mistral AI Data Processing Addendum
low
Added table of contents and section headers to Data Processing Addendum covering data definitions, security, breach procedures, subprocessing, and audit rights
Why it matters: The restructured Data Processing Addendum provides explicit section organization that clarifies which obligations appear in which section, making it easier for commercial customers and compliance officers to locate and verify specific data protection commitments. The substantive terms governing security, breach notification, subprocessing, and audit rights remain in place; this change primarily improves document usability without altering the scope of data protection obligations.
Mistral AI
Mistral AI Additional Product Terms
low
Expanded Additional Product Terms with sections on data classification, export procedures, deletion process, termination, and custom model training responsibilities.
Why it matters: The updated terms provide greater procedural clarity around data export, deletion, switching, and custom model training responsibilities. The formal classification of customer data as Exportable or Non-Exportable may affect how downstream vendors and customers can access or receive data processed through Mistral's infrastructure, with implications for data portability commitments and contractual compliance.
Mistral AI
Mistral AI Commercial Terms
low
Added table of contents and navigation structure to Commercial Terms document.
Why it matters: This change improves the usability of the Commercial Terms document by adding a clear table of contents and section navigation. Customers reviewing the terms will find it easier to locate relevant provisions, but the underlying contractual obligations, restrictions, and rights remain identical to those stated in the August 5, 2026 version.
Mistral AI
Mistral Terms of Use
low
Reorganized Legal Center landing page structure to improve document navigation and add introductory overview.
Why it matters: This change improves navigation and user experience for accessing Mistral AI's legal documents but does not modify the substantive rights, obligations, or policies themselves. Users seeking specific terms of service, privacy policies, or commercial agreements will find them through reorganized navigation rather than the previous structure.
Mistral AI
Mistral AI Usage Policy
low
Reorganized Usage Policy navigation and added language selection options; scope and substance unchanged.
Why it matters: This change improves documentation accessibility and organization without modifying substantive policy. The addition of language selection options makes the Usage Policy more accessible to non-English-speaking users, and the reorganized navigation helps users locate specific policy sections more easily.
Mistral AI
Mistral AI Terms of Service
low
Added organizational structure to Terms of Service with 14-section table of contents
Why it matters: This change improves the organization and navigability of the Terms of Service by adding a 14-section table of contents, making it easier for users to locate specific provisions. However, no substantive changes to consumer rights, data handling, liability, or dispute resolution procedures are evident from the structural reorganization.
Mistral AI
Mistral AI Privacy Policy
low
Restructured privacy policy with added table of contents and navigation elements.
Why it matters: This change improves the usability and accessibility of Mistral AI's privacy disclosures by reorganizing the policy into clearly labeled sections and adding navigation features. The substantive privacy terms remain unchanged, so this does not alter obligations or disclosures; it makes existing information easier to locate and review.
Google Gemini
Gemini 3.1 Pro Model Card
low
Updated model card references; Gemini 3.8 Flash added to latest model listings.
Why it matters: This change reflects Google's documentation of available model variants in its model card. It does not alter the terms of service, governance frameworks, or obligations for users, but it does indicate which model versions Google is currently promoting or feature in its product documentation.
Kindle
Kindle Store Terms of Use
low
Kindle Store Terms of Use change detected, but substantive policy text not accessible in provided diff.
Why it matters: The change detection indicates a modification to Kindle Store Terms, but the provided diff contains only page navigation and header content, not the actual policy language. Without access to substantive terms, operational significance cannot be assessed.
Peacock
Peacock Privacy Policy
low
Removed one service provider entity from German data protection contact information; reformatted section header.
Why it matters: The updated policy clarifies which NBCUniversal entity handles data protection inquiries for German users, reducing potential confusion and ensuring data subject requests reach the correct responsible party. Section header reformatting improves consistency in policy presentation.
Uniswap
Uniswap Privacy Policy
low
Updated last-modified timestamp in privacy policy metadata; no substantive policy changes.
Why it matters: This change has no operational significance. It is a routine metadata update to the privacy policy's last-modified timestamp. No substantive privacy terms, data handling practices, or user rights were altered.
Stay ahead of the changes

Don't manually check every platform

Get alerts when policies change, before it affects you.

Updated daily. New changes added as detected.

← Newer Page 3 of 99 Older →