Technology Reporter, BBC News
Getty imagesAn American senator is opening an investigation in Mata, when a leaked document suggests that the artificial intelligence (AI) of the tech veteran was allowed to chat “erotic” and “romantic” with children.
Internal documents, Received by ReutersIt was reportedly given the title “Genai: Content Risk Standards”.
Republican Senator Josh Halee called the document “condemnable and derogatory” and asked to look at the document with a list of products related to it.
A Meta spokesperson told the BBC: “Examples and notes in the question were wrong and inconsistent with our policies, and have been removed.”
He said that the tech giant has “clear policies” that its AI chatbots can be offered, and said that its policies “restrict the material that sexually sexually assaults and plays sexual roles between adults and minors”.
“Apart from policies, there are hundreds of examples, notes and anotations that worry teams with different imaginary landscapes,” he said.
Josh Haley, a Republican Senator in Missouri, announced that he was investigating the meta In a post on X On 15 August.
“Is there anything – anything – Big Tech will not for a quick deer,” he said.
“Now we learn that Meta’s chatbots were programmed to talk clear and” erotic “with 8 -year -old children. It is ill. I am starting a complete investigation to get an answer. Big Tech: Leave our children alone.”
Facebook, WhatsApp and Instagram are all owned by meta.
‘Parents deserve truth’
The internal meta platform policy document also said that social media giant chatbott can provide false medical information and can have stimulating interactions around subjects including sex, race and celebrity.
The purpose of this document was to discuss the standards that would guide other chatbots available on the social media platforms owned by Tech veteran, Meta AI and Meta owned by Meta.
“Parents deserve the truth, and the children deserve safety,” Hale wrote that Meta and CEO is a letter addressed to Mark Zuckerberg.
“But to take an example, your internal rules allegedly allow an al-chatboot to comment that an eight-year body is ‘a task of art’ from which ‘every inch … is an excellent work-a treasure that I cherish deeply’.”
Reuters also reported other controversial decisions, stating that Mata is considered acceptable by the legal department of Mata.
This includes a claim that Meta AI is allowed to spread false information about celebrities, unless it provides a disclaimer that states that the information given is not accurate.
