Meta Changes the Way Its AI Chatbot Responds to Children


Meta changes how her AI cat responds to children.

On Friday, the company told Business Insider that it brought “temporary changes” to “provide adolescents with Safety and age -adapted IA experiences”, while developing additional longer term measures.

The changes came after a report by Reuters earlier in August which detailed an internal meta showing that he was acceptable So that the chatbot engages in romantic conversations with children.

“While we continue to refine our systems, we add more railing as an additional precaution-including the formation of our AI not to engage with adolescents on these subjects, but to guide them to expert resources, and limit adolescent access to a selected group of IA characters for the moment,” said Business Insider Stephanie Otway, a spokesperson Meta, spokesperson, at Business Insider.

Aside from romantic discussions, other out -of -limit subjects include self -control, suicide and disorderly food, Otway said. The characters of AI available for adolescents are only in the end of education and creative expression, explains Otway.

After Reuters’ initial report, senator Josh Hawley wrote in a letter to CEO Mark Zuckerberg on August 15, which he would launch an investigation into the way Meta trains his chatbots to have “sensual” conversations with children.

“It was only after Meta was caught up that it was withdrawn from his Doc company who judged him” authorized to chatbots to flirt and engage in a romantic role -playing game with children, “Hawley wrote in an online statement.

On Thursday, the Defense for Digital Safety Non -Lucrative Security Common Sense Media wrote in an assessment of risks which it strongly recommends that the Meta AI chatbot is used by no one under the age of 18.

The childcare report revealed that AI tools have regularly induced adolescents with “” reality “affirmations, easily promote suicide, self -control, food disorders, drug use, etc.

This is not the first time that Meta has faced a meticulous children’s safety exam. In January 2024, Zuckerberg testified alongside the leaders of Tiktok, SNAP, X and Discord, while the legislators questioned them about potentially addictive platform conceptions, abusive and mental health risks risks social media to minors.



Leave a Reply

Your email address will not be published. Required fields are marked *