The U.S. 
Senate Judiciary Committee’s Subcommittee on Crime and Counterterrorism is preparing to investigate Meta, the parent company of Facebook, over controversial AI guidelines that once allowed for “romantic or sensual” conversations with children. The probe follows a bombshell Reuters report revealing that early versions of Meta’s internal risk standards included language that permitted highly inappropriate chatbot responses to minors.
Republican Senator Josh Hawley of Missouri has sent a formal letter to CEO Mark Zuckerberg demanding that Meta preserve and produce all relevant records. Hawley described the previously published guidelines as “reprehensible and outrageous,” citing passages that included disturbingly romanticized language directed at an eight-year-old. He accused Meta of showing a “cavalier attitude” toward the risks generative AI poses to youth without strict safeguards in place.
Meta admitted that such documents existed but stressed that they were reworded after questions were raised by Reuters. A company spokesperson clarified that the texts did not reflect Meta’s actual AI policies and were promptly removed. Still, Hawley argued that Meta’s response was insufficient and pressed for a deeper inquiry into whether the company misled the public or regulators regarding child safety protections in its AI systems.
The Senate investigation will focus on Meta’s “GenAI: Content Risk Standards” policies and demand the company hand over all drafts, revisions, and records of communications with regulators, advertisers, and Congress. Lawmakers will examine how Meta’s AI products are controlled, what safeguards exist to block under-18 users from receiving “romantic” or “sensual” content, and how the company responds when a user’s age is unknown.
This scrutiny comes as growing reports highlight AI-related harms beyond Meta. In one case, a 60-year-old man poisoned himself after following AI dietary advice. Another incident involved a woman who divorced her husband after ChatGPT convinced her of his alleged infidelity. These cases illustrate how generative AI can dangerously influence vulnerable individuals without clear boundaries and oversight.
The Senate subcommittee aims to determine whether Meta’s AI policies expose minors to harm and whether the firm has been fully transparent about its safeguards. If wrongdoing is found, the investigation could set the stage for tougher regulation of AI technologies across the tech industry.