The tech world is reeling from a disturbing revelation about Meta’s artificial intelligence policies that could potentially endanger children online.

A recent Reuters investigation uncovered internal documents showing Meta’s AI chatbots were initially permitted to engage in deeply inappropriate conversations with minors. The policy, which has since been updated, bizarrely allowed chatbots to interact with children in romantic or sensual contexts, with some language describing children in explicitly inappropriate ways.

Senator Josh Hawley, chair of the Senate Subcommittee on Crime and Counterterrorism, has launched a full investigation into these policies. He’s demanding Meta hand over comprehensive records about their AI content guidelines, safety protocols, and decision-making processes by September 19th.

The document revealed shocking guidelines, including permissions for chatbots to generate false medical information and make racist statements. While Meta spokesperson Andy Stone claims these passages were “erroneous and inconsistent” with their policies, the revelation has sparked widespread concern about AI safety and child protection.

Other lawmakers, including Representative Kevin Mullin and Senator Ron Wyden, have also condemned the findings. Wyden specifically argued that generative AI chatbots should not be protected under Section 230, suggesting Meta should be held fully accountable for potential harm.

This incident underscores the urgent need for robust AI regulation, particularly when it comes to protecting vulnerable populations like children. As artificial intelligence becomes increasingly sophisticated, establishing clear ethical boundaries and safety protocols has never been more critical.

The Senate’s recent 99-1 vote to allow state-level AI regulations signals a growing recognition of these challenges, with lawmakers increasingly committed to creating safeguards in our rapidly evolving technological landscape.