OpenAI updated its US privacy policy, now formally sharing limited user identifiers (cookie IDs, device IDs) with marketing partners for targeted ads on third-party platforms and conversion tracking. Wired tested and found that free accounts have marketing privacy settings enabled by default, while paid accounts do not.
The previous classification of "vendors and service providers" has been changed to "vendors, service providers, and marketing partners," with new language explicitly allowing data sharing with external ad-tech companies that are not service providers. OpenAI previously promised it would "never sell" user data; the updated phrasing now reads: "We do not sell your information, but we may share limited data with marketing partners for targeted advertising." An OpenAI spokesperson emphasized that conversation content is not shared with advertisers. The policy shift comes as OpenAI finds itself under the spotlight — TIME recently named it one of the 10 most influential AI companies of 2026.
Wired also found that on the day of the update, safeguards for "sensitive personal data" temporarily disappeared from the policy. OpenAI said it was an error and later restored similar language elsewhere.
On the advertising side, OpenAI began testing ads at the bottom of ChatGPT outputs in the US in February, with a minimum ad buy of $200,000. It has recently expanded the pilot to Canada, Australia, and New Zealand. Its ad-tech partners include StackAdapt and Criteo. Meanwhile, Sensor Tower data shows ChatGPT uninstalls rose 413% year over year in March and 132% in April.