Key figure: 63% of organizations in IBM's 2025 breach study had no AI governance policy, and 97% of AI-breach victims had no proper AI access controls (IBM, Cost of a Data Breach Report 2025).
This cybersecurity news time machine tracks AI vendor terms of service changes over time. OpenAI, Anthropic, Meta, Google — see exactly what each vendor changed and when.
It tracks every significant change to major AI vendors' terms of service: data-training clauses, liability caps, retention windows, and pricing terms. Each entry shows the before-and-after wording with dates, so you can see the direction vendors are moving in 2026.
Pick an AI vendor. Scroll its timeline. Every entry in this terms of service changes tracker is a real, dated, sourced change to its terms or privacy policy — and for each one, we say what they took from you. Where we can't verify a change, we say that too: an honest gap beats an invented entry.
TermsWatch monitors vendor terms and privacy policies so you catch the next change the day it lands — not the day after your data is gone. And the Security Suite bundles every Ingenious AI platform in one early-access package.
Our platforms are being upgraded and return November 1 — get early access now:
Get TermsWatch early access →In the October 17, 2026 check, Anthropic announced a usage-policy overhaul effective November 12, 2026, adding conduct rules for users, and Meta removed the word "train" from its commercial terms while continuing to crawl business sites for ads and business products. OpenAI, Google, and xAI showed no verifiable change that week. Dates and sources are on the timeline above.
In the check run of October 17, 2026: Anthropic announced a usage-policy overhaul, and several vendors adjusted data-training clauses.
It depends on your plan's settings — and terms change, which is why the timeline exists. Check your current plan's data controls.
The original commercial-terms clause took business-linked site content to train AI products — later narrowed after scrutiny.
TermsWatch diffs vendor terms and privacy policies so deletions and quiet weakenings surface before they cost you.
In the check run of October 17, 2026: Anthropic announced a usage-policy overhaul (effective November 12, 2026) and Meta removed the AI-"train" sentence from its commercial terms. OpenAI, Google, and xAI had no verifiable change that week.
It depends on your plan's settings — and terms change, which is why the timeline exists. In the October 17, 2026 check, OpenAI had no verifiable change: its usage-policy changelog tops out at October 29, 2025. Always verify against the current policy.
The original commercial-terms clause took business-linked site content to "train" AI products with no opt-out. After October 17, 2026, the word "train" is gone — but broad crawling for ads and business products remains, and Meta says its crawlers operate as before.
TermsWatch diffs vendor terms and privacy policies so deletions and quiet weakenings surface the day they land. Get early access via the WhatsApp link above.