OpenAI removes ChatGPT's public sharing feature after Google exposes sensitive chats

OpenAI has removed ChatGPT 'Make this link discoverable' feature after several users discovered that a lot of their sensitive shared conversations became visible through Google Search. The "feature" was prompted by a simple checkbox when creating a shareable link and allowed conversations to be indexed if enabled, but many users mistakenly assumed their shared links would only be seen by direct recipients.

Following public reports of exposed data, OpenAI’s Chief Information Security Officer Dan Stuckey confirmed the feature was disabled late last week. The company also removed both the discoverable checkbox and all previously indexed chats from Google. However, cleanup is still underway, and some shared conversations remain accessible through Microsoft Bing and DuckDuckGo as the process continues. For now, no ChatGPT share links appear in Google results.

The incident revealed that exposed chat content included mental health disclosures, job applicant reviews, proprietary code, and even crime confessions. Search engine crawlers may have already preserved some content, making complete deletion challenging. This event raises a serious privacy concern, similar to the recent Meta AI app case where private user conversations were shared publicly without their full awareness.