Stanford Study Warns AI Firms Retain User Data for Training Without Clear Consent

Published
Score
17

Why it matters

Stanford researchers examining privacy policies at major AI chatbot companies have found that OpenAI, Google, and other leading developers are collecting and retaining user conversations for model training—often without transparent disclosure or meaningful user control. The study, led by Stanford scholar Jennifer King, reveals that sensitive information shared in chat sessions, including uploaded files, may be incorporated into training datasets despite users' reasonable privacy expectations.

The research identified several specific practices raising concern: extended data retention periods, collection of children's data without adequate safeguards, and opaque privacy disclosures. Google has announced plans to train models on teenage users' data if they opt in. Anthropic bars users under 18 but does not require age verification. Currently, AI companies operate under a patchwork of state-level laws and lack meaningful federal regulation, while the Fourth Amendment protects private communications from government searches but does not apply to private commercial platforms.

Attorneys should monitor this issue closely as it sits at the intersection of commercial surveillance and potential government access to chat data. Unless users affirmatively opt out, their conversations become training material. Cybersecurity experts have begun arguing that AI prompts deserve Fourth Amendment protection and that companies should resist bulk government data requests without warrants. As AI adoption expands across industries, expect increased regulatory scrutiny and potential litigation over how platforms handle sensitive user information.

mail

Get notified about new Privacy developments

Primary sources. No fluff. Straight to your inbox.

See more entries tagged Privacy.

Also on LawSnap