A ChatGPT bug caused a privacy breach that exposed private chat histories of some users to others. As a result, numerous users took to various social media and online forums to share screenshots of chat conversations that did not belong to them. OpenAI, the developer of ChatGPT, acknowledged the glitch to Bloomberg, clarifying that only the conversation titles were displayed, not the complete conversations themselves.
In response to this issue, OpenAI’s CEO, Sam Altman, took to Twitter, disclosing that the cause of the bug was a technical glitch in the open-source library. He expressed regret for the situation, saying that a “technical postmortem” would follow soon. Upon discovering the bug, OpenAI disabled chat histories on Monday which was brought back later the same night.
This privacy breach highlights the potential risks of using chatbots and AI-based systems, and the significance of not sharing private details on such platforms. OpenAI, on its FAQ page writes warns users not to share “any sensitive information in your conversations” saying that it cannot share specific prompts from chat history. Users trying to access chat history during the outage saw a message reading ““History is temporarily unavailable” and that the company is working to restore the feature as soon as possible.
While OpenAI says that the service has been restored now, Altman informs that users won’t see their histories chats of a specific time period from March 27.
ChatGPT chat history appears on the left hand side of the website. As the name suggests, it lets users view the past conversations. While it has no role in starting a new conversation, it is needed in case a user wants to pick up an old conversation and continue with it.
ChatGPT is an AI-powered chatbot from OpenAI built to give human-like responses. It shot to fame and has garnered a huge user base due to its conversational style of interacting with users.
