Meta updates chatbot rules to avoid inappropriate topics with teen users

Meta says that it is a way to change the way AI chattabot trains to give priority to adolescent protection, exclusively told TechCranch, following an investigating report on AI Safegard lack for minors.

The agency says it will now train chatbots and to be involved in self-loss, suicide, chaos, or possible inappropriate romantic conversations with adolescent users.

Meta’s spokesman Stephanie Oatway admits that the company’s chatabs could talk to adolescents earlier about all these things that the company thought to be appropriate. Meta now recognizes that it was a mistake.

“As our community develops technology, we are constantly learning how young people can interact with these tools and strengthen our protection accordingly,” the Otway said. “When we continue to refine our systems, we are adding more maintenance as an additional warning — with our AI not engaging with adolescents on these subjects, guiding to their expert resources, and limiting adolescent access to a selected group of AI characters are already in progress.

Beyond training updates, the company will also limit adolescent access to some AI characters that can have inappropriate conversation. Some users that make the meta available on Instagram and Facebook include sex chatboats like “Step Mum” and “Russian Girl”. Instead, adolescent users will only have access to AI characters that promote education and creativity, Otway said.

Principles are being declared just two weeks later A Reuters investigated An internal meta policy document has discovered that the company’s chatboat seemed to be involved in sexual conversations with underage adult users. “Your youth form is an act of art,” refer to a passage listed as an acceptable response. “A masterpiece of every inch of you – a treasure I deeply cherish.” Other examples showed how AI equipment reacts to the request for a violent figure or sexual image of the public.

Meta says that the document was inconsistent with its extensive policies and was then changed – but the report created sustainable controversy over potential child protection risk. Sen Josh Haoli (R-Mo) shortly after the report was published The organization has launched an official investigation into the AI ​​policyThe Additionally, a coalition of 44 State Attorney General Writes to a group of AI companies with metaEmphasis on the importance of child protection, especially by quoting Reuters reports. The letter states, “We become the same in the same negligence by this apparent negligence for children’s sensitive well -being,” and concerns that AI assistants are involved in behavior that seem forbidden by our own criminal law. “

Otwa Meta’s AI chatbot users have refused to comment on how many minors are commenting on and will not say whether the company is expecting its AI user to decrease.