Meta Pulls the Plug on Teen-Chat AI Characters Days Before Child-Safety Trials

Silicon Valley has been here before: in 1997 America Online shut down its “Teen Chat” rooms after high-profile predator cases, and Congress responded with COPPA in 2000. Today’s generative-AI chatbots reprise that cycle—new conversational tech, old worries about grooming and self-harm. Meta’s pre-trial shutdown looks less like sudden altruism than a classic legal risk-mitigation tactic, echoing Big Tobacco’s last-minute advertising curbs before the 1998 Master Settlement. Long-term, the pattern is that each new medium (radio in the 1930s, TV cartoons in the 1980s, social media in the 2010s) expands rapidly until child-protection politics force guardrails; the guardrails then become industry baseline. Whether Meta’s AI companions return in a tamer form or disappear altogether will signal how generative AI products must adapt over the next century to survive increasing demands for age gating, transparency and legal liability.