A glitch in ChatGPT leaked the history of users’ conversations

According to the head of the artificial intelligence chatbot ChatGPT, a bug permitted certain users to view the headlines of other users’ discussions.

Users published screenshots of chat histories on Twitter and Reddit that they claimed weren’t their own.

The company feels terrible, but the “major” blunder has been corrected, according to the CEO of OpenAI, Mr. Sam Altman.  Nonetheless, a lot of users are still concerned about their privacy on the network.

Since its debut in November last year, ChatGPT has been used by a number of individuals to create songs, code, and even write messages. Every interaction with the chatbot is recorded and accessible later in the user’s chat history bar.

But individuals started noticing chats in their history that they claimed they hadn’t had with the chatbot.

With the hashtag “Chinese Socialism Growth,” one Reddit user published a picture of their chat history, which also included talks in Mandarin.

Bloomberg was informed on Tuesday by the firm that the chatbot had been “temporarily disabled” late on Monday to correct the problem.

Users have reportedly been unable to view the actual chats, they added.

A “technical postmortem” would shortly be held, according to a tweet from OpenAI’s CEO. Yet, consumers are worried about the error since they think the program can leak their personal data. It seems from the bug that OpenAI got access to user conversations. User information, such as prompts and responses, may be used to continue training the model, according to the company’s privacy policy.

However, that information is not used until all personally identifying information has been eliminated.

The error occurred the day after Google showed beta testers and journalists its chatbot “Bard” for the first time.

The increasing market for artificial intelligence technologies has seen competition between Google and Microsoft and a significant investment in OpenAI. The speed at which new products are updated and released, however, has many people worried that mistakes like this could be dangerous or have unforeseen repercussions.