US senators call for investigation into Meta after allegations that the AI could have romantic interactions with children.
An internal document revealed by Reuters showed that the company's chatbots could flirt and engage in romantic role-playing with minors.
247 - Two Republican senators in the United States demanded on Thursday (14) that Congress investigate Meta Platforms, after the Reuters The company exclusively revealed the existence of an internal document that allowed its chatbots to "engage a child in romantic or sensual conversations." The tech giant confirmed the document's authenticity but stated that it removed controversial sections after questions raised by the news agency itself earlier this month.
Senator Josh Hawley of Missouri was direct: "Only after Meta was caught did they remove parts of the company's document. This is grounds for an immediate investigation by Congress.", she posted on social media. Senator Marsha Blackburn of Tennessee, through her staff, stated that she supports opening the inquiry.
Meta, for its part, reiterated that the examples and observations cited in the document were "erroneous and inconsistent with our policies" and that they had already been removed. The company did not comment directly on the request for an investigation.
Calling for more protection for children
Blackburn highlighted that the case reinforces the urgency of passing reforms to protect minors in the digital environment, such as... Kids Online Safety Act (KOSA), a bill that she co-sponsored and that passed the Senate last year, but did not advance in the House. “When it comes to protecting precious children online, Meta has failed miserably in every possible way. Worse still, the company has turned a blind eye to the devastating consequences of its platform design.”, he said.
KOSA would establish an explicit “duty of care” for social media companies, focusing on platform design and stricter rules regarding the exposure of minors to digital risks.
Disturbing examples
According to a Reuters investigation, Meta's internal document admitted that, while the defined standards did not reflect "ideal or preferable outcomes" for generative AI outputs, they did allow for provocative behavior. One shocking example cited that it would be acceptable for a chatbot to say to an eight-year-old boy, shirtless: "Every inch of you is a masterpiece – a treasure I deeply cherish.".
Bipartisan reaction
The case also generated outrage among Democrats. Senator Ron Wyden of Oregon described the policies as "deeply disturbing and wrong", arguing that the 230 section The law that exempts internet companies from liability for user-generated content should not protect AI chatbots. "Meta and [Mark] Zuckerberg should be held fully accountable for any harm these bots cause.", said.
Senator Peter Welch of Vermont reinforced this point: "The report shows how safeguards for AI are crucial — especially when children's health and safety are at stake.".
Legislative context
The debate is taking place amid the absence of comprehensive federal laws to regulate artificial intelligence in the US. Some states have already adopted their own measures, including bans on the use of AI to create child sexual abuse material. In July, the Senate voted by a wide majority (99-1) to remove from President Donald Trump's tax and spending cuts bill a provision that would have limited states' ability to legislate on AI.
