ChatGpt conversations indexed on Google: here's the Ai that 'spies' on us

You feel like learning the art of prompting if the content of ChatGpt chats ends up on the Net and becomes available to anyone. It is not an impromptu provocation that leaves time to find the one raised by the FastCompany.com portal, publishing an article from which it can be deduced how it is possible to retrieve on Google some 4500 conversations between anonymous users and OpenAi's generative chatbot. The discovery concerns interactions in several languages and on the most disparate of times, ranging from strictly topical subjects to even sensitive issues such as very serious health problems, but the question is above all another one: how can BigG's search engine index these conversations if the entire flow of questions and answers is secreted and no one (not even OpenAi) can access users' chats? What more or less everyone knows is that Sam Altman's creature (and so do the other generative intelligences) can anonymously use the content of these conversations to improve AI models, and only if the user chooses to contribute to the model training process. What then?

The secret revealed by FastCompany is actually a technical procedure, obviously unknown to most. In a nutshell, if one puts a tick in the box that enables the command to 'make chat searchable', it becomes content that can be displayed in web searches, and therefore indexable on search engines. OpenAi, and this is a matter of public knowledge, has for some time been offering users the possibility of sharing ChatGpt dialogues through a system that generates a link (containing only the contents of the conversation and not the personal data of the person who generated it) that can be shared by anyone who comes into possession of it. How can one avoid putting one's interactions with AI out in the open? The answer is much simpler than you might think: just leave this function deactivated as it is by default, and thus avoid selecting it and essentially removing the veil of privacy from your conversations.

OpenAi, for its part, has decided to deactivate the function that allowed users to make their conversations indexable by search engines, and as its Chief Information Security Officer, Dane Stuckey, wrote on X, the reason is quickly explained: 'we felt that this function introduced too many possibilities for users to accidentally share unwanted content. We are removing the option and working to remove already indexed content from search engines'. The GenAi saga, meanwhile, has known another chapter.