OpenAI recently confirmed that they suffered a privacy breach that resulted in ChatGPT leaking conversation histories of random users.

ChatGPT will display your past conversations with AI on the left side of the interface, and save the content in the system, so that you can review the previous content from time to time. It’s just that some users recently discovered that their chat history shows conversations that never happened, and these contents seem to come from other users. Meanwhile, ChatGPT experienced some glitches on the same day, which left users unable to access.

The incident happened on Monday, but OpenAI did not respond until Wednesday, confirming that the content of user conversations would unexpectedly appear in other users’ histories.

After this problem occurred, many users were worried whether the problem was caused by hackers. A few days ago, Open AI CEO Sam Altman personally postedtweetsexpress:

We had a major issue in ChatGPT due to a bug in an open source library, and now a fix has been released and we’ve just completed verification.

A small percentage of users are able to see the headings of other users’ conversation histories.

At present, OpenAI has not announced which open source library has the problem, and we still don’t know whether the vulnerability leaked the user’s personal information or the credit card and other payment information of ChatGPT Plus users in addition to leaking the conversation record.

PS:Regarding this vulnerability, I also asked ChatGPT at the end, and it didn’t help me find the answer (it may also be because I am not a valued ChatGPT Plus member and do not have the latest data ☹️).