Feature, not a bug: OpenAI kills ChatGPT public chat search after users overshared weird, personal stuff

India Today

4 days ago

OpenAI has decided to take down a feature that allowed ChatGPT conversations to appear on Google Search, after concerns were raised over user privacy. The feature was meant to help people discover useful conversations, but it ended up creating problems as some personal chats started showing up publicly.The update, rolled out earlier this year, gave users the option to share specific ChatGPT conversations in a way that made them searchable on platforms like Google. According to OpenAI, this was an opt-in feature and not turned on by default. However, many users were not fully aware that enabling this option could make their chats appear in search results, and some accidentally shared private details without realising the consequences.advertisementOpenAI's Chief Information Security Officer, Dane Stuckey, announced the decision to remove the feature through a post on X (formerly Twitter). He explained that while the idea was to help people access helpful public content, it also led to situations where users unknowingly shared things they didn't mean to.
"We just removed a feature from @ChatGPTapp that allowed users to make their conversations discoverable by search engines, such as Google. This was a short-lived experiment to help people discover useful conversations. This feature required users to opt in, first by picking a chat to share, then by clicking a checkbox for it to be shared with search engines (see below)," Stuckey wrote.He clarified that the sharing process involved selecting a chat and then ticking a box to allow it to be searchable, but even then, the risk of accidental exposure was too high. As a result, the company has now removed the option entirely and is working with search engines to take down any chats that have already been indexed.The issue gained attention after it was discovered that thousands of ChatGPT conversations had become visible on Google. A report by Fast Company revealed that over 4,500 chats were publicly searchable. While many of them were general and harmless, some included personal information such as names, locations, or sensitive thoughts shared by users in confidence.What made the situation more concerning was that deleting a chat or its link didn't instantly remove it from search engines. These chats could still show up online until Google and others updated their search index, which takes time.OpenAI CEO Sam Altman also addressed the matter in a recent podcast. He pointed out that users often open up to ChatGPT and discuss private matters, trusting the AI with details they may not even share with others. This level of openness is exactly what made the sharing feature risky, despite being optional. - Ends