OpenAI has just removed a feature that once made ChatGPT dialogue sections shared publicly appear on Google search results.
This was originally a short-term experiment based on the option of creating shared links. After a series of user reactions, Mr. Dane Stuckey - Director of Information Security at OpenAI - said that the company is working with search engines to remove these chats from Google and similar platforms.
The story originates from the fact that users discover many ChatGPT dialogue sections that can be found through Google. Although these chats do not clearly state the user's name, some content still contains sensitive information that can reveal the identity of the author.
It should be emphasized that this is not a leak or data attack. The problem lies in an available option when users create a link to share the chat. In the window that appears, there is a check box with the words "Allow this chat section to be found". A more specific explanation allowing them to appear in web search results is placed smaller, less clear below.
Users are only affected if they check this box. However, many people may have mistaken or just want to save the link to send via text message or reopen later, not expecting that the content will be indexed by Google, that is, collected and displayed publicly by the system on the network.
In his initial response, Stuckey said the description was clear enough. But after a wave of criticism, OpenAI changed its decision. We found that this feature allowed users to accidentally share unwanted information, so we decided to completely remove that option, said Stuckey.
