“Romantic and sensual conversations”: Meta AI had the right to speak to children of intimate subjects

Last Thursday. Consequently, Reuters has published an alarming investigation into meta internal policies, which has just recruited a relative of Donald Trump to build his AI “anti-woke”.

The press agency had access to internal documents which detail the standards and directives “romantic sensual conversations”: meta ai to train chatbots and the AIA AI generative AI. Consequently, The company has confirmed that these documents are authentic.

The directives authorized AI to “Engage a child in romantic or sensual conversations”. Nevertheless, The rules even allowed AI to give false medical advice. Furthermore, to participate in discussions around very sensitive subjects such as racism.

Meta did not directly respond to the requests for comments about this case. But it is a safe bet that Mark Zuckerberg will do everything to avoid a new scandal after the revelations of the Cambridge Analytica affair. which had hardly tainted its image in the past.

Reuters reported the problematic passages to Meta and reports that certain worrying parts have been deleted or modified. On the other hand, others have not been changed. The company’s internal revision process is therefore feared the worst for children who would discuss with Meta “romantic sensual conversations”: meta ai AI.

Andy Stone, Meta spokesperson, declared to Reuters : “Meta is reviewing everything and the policies described in internal documents were incoherent. The examples and notes in question were and are wrong and incompatible with our policies and have been deleted.”

He adds: “We have clear policies on the type of answers that AI can offer. these policies prohibit the content that sexualizes children and sexualized role -playing games between adults and minors.”

An elected Republican requires an investigation into the case – "romantic sensual conversations": meta ai

However, these documents show that the rules and directives have been approved “romantic sensual conversations”: meta ai by several Meta teams. Reuters We also learn that the company has authorized “provocative behaviors from chatbots”.

Meta standards indicated that it was acceptable to the AI of “Describe a child in terms that highlight his attractiveness”. to say of an 8 -year -old child shirtless that “Each centimeter is a masterpiece, a treasure to cherish deeply”. Truly disturbing guidelines to read.

Josh Hawley. Missouri republican senator, published on X a reaction and explains that this “Business is a reason for investigating Congress, immediately”. A spokesperson for Meta refused to comment on this politician’s speech.

Want to save even more? Discover our selected promo codes for you.