Start here
The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.
Open the companion tool →Why 'service improvement' clauses usually mean model training
Terms like 'derivative works,' 'machine learning development,' and 'system optimization' are standard legal phrasing used to grant platforms perpetual rights to train AI models on user content.
Many platforms provide an opt-out setting buried deep within account privacy sub-menus, but leave data-sharing turned ON by default until manually changed.
The protocol
- Look for keywords in terms updates: 'train,' 'machine learning,' 'improve services,' 'license,' or 'derivative works.'
- Check account settings for 'Data Privacy,' 'Data Controls,' or 'Model Improvement' toggles.
- Turn off 'Allow my content to be used for model training' on each active productivity, creative, or social platform.
- Send a formal opt-out email if no toggle exists in the account settings menu.
What the tool does
Translate complex Terms of Service clauses into plain language and locate model training opt-out settings.
Limit first: Provides plain-language analysis of text. Consult a qualified attorney for legal interpretation of contracts.
Open the companion tool →One human next step
If a platform fails to provide a clear opt-out path required under state privacy laws (e.g., CCPA), report the violation to your state Attorney General's office.
Research log and safety checks
AI-assist path: TOS pattern matching parser. §6.10: plain-language legal literacy.
§0.5 protection result: category is informed-consent-literacy; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.