Normally, ChatGPT users can see the history of their conversations with the AI. The problem is that on Monday, users started complaining about being shown snippets of conversations they definitely didn’t participate in, to the point where one of the users saw a history of Chinese conversations in the sidebar.
Image Source: Rolf van Root/unsplash.com
Users fear that their accounts may 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 simply “invents” the previous fragments.
OpenAI reported ChatGPT outages on Monday. At the same time, on the contrary, it was reported about the loss of chat histories by users, and not about the discovery of foreign fragments. One way or another, according to OpenAI, the performance and conversation histories have been restored.
Image Source: Reddit
User complaints once again raise the issue of the safety of using such services. It is hardly worth entering meaningful information, knowing that it may be available to other users. The good news is that the users reporting the issue stated that it was only showing the “titles” of previous chats, not the full-text materials – if you click on such a title, the program will not be able to download the entire conversation. In addition, users did not see someone else’s personal information – only common phrases, although some on the ChatGPT Plus payment page appeared data about unknown postal addresses.
In other words, ChatGPT, like any web service, can potentially suffer from a data breach. In addition, the OpenAI privacy policy provides for the transfer of some aggregated information such as user statistics to third parties and even the publication of such information. At the same time, ChatGPT in the “questions and answers” reminds that confidential information should not be entered during conversations.
If you notice an error, select it with the mouse and press CTRL + ENTER.
