OpenAI CEO Sam Altman expressed regret after a bug in the company’s popular AI tool ChatGPT caused a leak of users’ chats. The issue allowed some ChatGPT users to view the titles of other users’ conversations within the tool. OpenAI has since implemented a fix for the bug.
The problem originated from a bug in an open-source library. Altman did not disclose the name of the open-source library or provide details on how it had been utilized by OpenAI.
ChatGPT has experienced rapid growth since its launch late last year. Users worldwide have utilized prompts to generate creative content ranging from poems and novels to jokes and film scripts using the conversational chatbot.
Recently, Microsoft Corp-backed OpenAI introduced its latest artificial intelligence model GPT-4. This upgrade from GPT-3.5, which was released to users in November, has demonstrated significantly greater capabilities than its predecessor. GPT-4 can quickly execute complex prompts including completing tax returns and generating intricate sentence structures.
During its launch demonstration, OpenAI showcased ChatGPT-4’s ability to interpret images. The AI tool was able to create a website based solely on a hand-drawn sketch of its design.
The integration of OpenAI’s GPT technology into Microsoft’s Bing search engine has increased traffic to the platform according to data from analytics firm Similarweb.
Earlier this month tech directors from National Cyber Security Centre cautioned users against sharing personal or sensitive data with ChatGPT and other Large Language Models. This warning was issued due to concerns over potential malicious or accidental leaks as well as the possibility that developers may use chat data when training programs in future .
