Usually, ChatGPT users can see the history of their conversations with the AI. The problem is that on Monday, users complained about being shown snippets of conversations they definitely hadn’t participated in, to the point where one of the users saw a history of Chinese conversations in the sidebar.
Users fear their accounts have been hacked. According to the PCMag portal, no response has yet been received from the ChatGPT team, so it is not clear whether we are really talking about the history of real chats, or the program is simply “inventing” the previous fragments.
Open AI reported crashes in ChatGPT on Monday. At the same time, on the contrary, the loss of chat histories by users, and not the detection of foreign fragments, was reported. One way or another, performance and conversation histories have been restored, according to OpenAI.
User complaints again raise the question of security when using such services. There is little point in entering meaningful information knowing that it may be available to other users. The good news is that the users who reported the issue reported that only the “titles” of previous chats were displayed, not the full-text materials – clicking on such a title prevents the program from downloading the entire conversation. In addition, users did not see anyone else’s personal information – only common phrases, although some data about unknown mailing addresses surfaced on the ChatGPT Plus payment page.