OpenAI has faced a wave of outrage after users discovered that their private chats via ChatGPT suddenly appeared in Google search results. The crisis began after a report published by Fast Company revealed that thousands of conversations were publicly available online, despite containing personal and sensitive details.

The report showed that some of the conversations contained accurate descriptions of personal relationships, experiences related to mental health, drugs, or private life, raising serious concerns that others could identify the authors of these messages even without mentioning their names.

How did the conversations get to the search engines?

OpenAI explained that the conversations that came out publicly were the result of users activating the conversation sharing option, then clicking on a box titled Make the conversation discoverable. However, reports indicated that this message appeared in small font and a faint color, which made many users unaware of the consequences of what they agreed to.

Some users used to share links to their WhatsApp conversations or save them to return to later, without knowing that they became available through Google search. While OpenAI confirmed that it was done based on users' consent, the way the option was presented sparked a wide debate about transparency and clarity.

The emergence of sensitive conversations on Google has caused widespread shock, and Carissa Velez, a researcher in artificial intelligence ethics at the University of Oxford, expressed her shock at the recording of these sensitive conversations in the search engine.

Google did not comment directly on the incident, but clarified that controlling the appearance of pages in search results depends on content publishers and not on the search engines themselves. It noted that it is up to the publisher to determine whether a page is indexable or not.