ChatGPT bug leaks user’s chat history: Report

OpenAI-created ChatGPT, which was launched in November, quickly become popular for its ability to respond to difficult queries or write sonnets or code. It even answered medical and legal exam questions set for human students and scored high marks.

However, it appears that it is not without risks as its learning system. According to a new report, some users’ previous chat history with the AI chatbot was exposed to other users of OpenAI’s ChatGPT. This reportedly happened due to a bug.

Following the incident, OpenAI temporarily shut down its ChatGPT service on Monday morning.

An OpenAI spokesperson reportedly revealed that the titles were visible in the user-history sidebar that typically appears on the left side of the ChatGPT webpage. He further added that the company temporarily disabled the chatbot after receiving reports of this bug. The context of the other users’ conversations was not visible.

OpenAI said that the bug/error appeared due to an undisclosed open-source software.

However, to the relief of users only the titles of past chats with the chatbot were visible, not the full text of the conversations.

Meanwhile, OpenAI released GPT-4, the next-generation version of the technology that powers ChatGPT and Microsoft’s new Bing browser, with similar safeguards. GPT-4 has showcased good performance in early tests. It can reportedly draft lawsuits, pass standardized exams, and build a working website from a hand-drawn sketch.