OpenAI has quietly rewritten how ChatGPT handles personal data for users in the US, and the new terms make one thing plain: free access now comes with a much broader data trade-off. The company says it can analyze chat history for ad targeting, retain more information in some cases, and use contact uploads in ways that reach beyond the person doing the uploading.
That is a pretty familiar Silicon Valley bargain, only this time the fine print is doing most of the talking. OpenAI is also drawing a sharper line between the US and the rest of the world, which suggests the company is trying to grow its ad business without running headfirst into stricter privacy expectations elsewhere.
Ads are now part of ChatGPT’s free tier
Under the updated agreement, people on the free tier and subscribers to ChatGPT Go, which costs $8 a month, may see ads informed by their chats and browsing behavior inside the product. OpenAI says advertisers do not get direct access to personal information or conversation logs; they see aggregated data on impressions and clicks instead. That is still advertising, just with a privacy layer thick enough to satisfy lawyers.
ChatGPT contact uploads can expose people outside the app
The new rules also allow users to upload address books in order to find other ChatGPT users, but that creates an awkward privacy spillover. OpenAI can access the contact details of people in those lists even if they never signed up for the service, and they have no way to opt out of being swept into the system.
That kind of graph-building is useful for a platform trying to look social, and it also gives OpenAI more data to map relationships between users. Meta, Google, and other ad-heavy platforms have spent years refining similar networks; OpenAI is arriving late, but it is arriving with the same incentives.
US users get different treatment from everyone else
OpenAI’s policy now varies by region. In the US, the company reserves the right to share some user data with marketing partners for advertising purposes, while it says no such transfers are planned for the rest of the world. The split is a tell: the company’s commercial ambitions are colliding with a patchwork of privacy rules, and the US is where it appears willing to push furthest.
The company also says it receives feedback from ad partners about purchases that follow a click on an ad, and that data is used to measure campaign performance. In other words, ChatGPT is being wired into the standard ad-tech loop, one that has powered every platform from search to social media.
How long ChatGPT keeps your data
OpenAI says user-initiated deletions are cleared within 30 days, and temporary chat sessions are removed automatically. But the policy includes broad exceptions: information tied to financial transactions, illegal activity, or abuse can be kept longer, and accounts flagged for policy violations may have data retained indefinitely for possible disputes.
There is also a separate storage setup for all user data between April and September 2025 because of the New York Times copyright lawsuit, with access limited to OpenAI’s security and legal teams. For businesses using ChatGPT, the direction is obvious enough: once advertising becomes part of the product, the service starts behaving less like a neutral assistant and more like a platform with its own commercial agenda.
The open question is how many users will accept that trade-off without moving up to pricier plans. If OpenAI keeps leaning on ads, enterprise customers may also have to think harder about what sort of assistant they are actually deploying.