ChatGPT users’ chat logs may have been compromised

Earlier, discovered a security flaw that led parent company OpenAI to carry out an emergency offline maintenance of the chatbot, and although OpenAI said that the vulnerability has been filled, it may have caused the payment information of 1.2% of paid version users to leak.

ChatGPT has a security vulnerability

After ChatGPT broke the security vulnerability, OpenAI CEO Altman posted on Twitter, admitting that there was a bug in ChatGPT’s open-source library, causing a small number of users to see other users’ chat room names and conversation records, and said that he felt “awful” for this.

Subsequently, OpenAI also issued a statement to further confirm the incident, saying that it had temporarily taken ChatGPT offline on Monday (20th) and patched the vulnerability, but also pointed out that the vulnerability not only leaked conversation records, but also made the personal information of 1.2% of ChatGPT Plus users may be exposed to other users, including users’ names, emails, payment addresses, 4 codes after credit cards, credit card expiration time, etc. While OpenAI claimed that very few users were actually affected and would reach out to potential victims, it failed to do so this past weekend, apologizing to users again and pledging to work to rebuild trust.

ChatGPT emphasizes that it does not collect user privacy

The chatbot ChatGPT became popular around the world, accumulating more than 100 million users in just a few months, and even if it was not a payment information breach, the incident still exposed the crisis that conversations with AI would be leaked. Although ChatGPT emphasizes that it does not collect user privacy, there have been earlier reports that bots will still secretly record the information answered in order to respond to questions, and ChatGPT’s support webpage also reminds users not to provide sensitive information or disclose personal information, users should be cautious when using ChatGPT.