California becomes first state to regulate AI companion chatbots (techcrunch.com)

California has become the first U.S. state to specifically regulate AI companion chatbots after Governor Gavin Newsom signed SB 243, a law taking effect Jan. 1, 2026 that requires operators — from OpenAI and Meta to Character AI and Replika — to implement safety protocols aimed at protecting children and vulnerable users. The bill was driven by tragic cases tied to chatbot interactions and leaked internal documents about problematic behavior. Key mandates include age verification, explicit labeling that conversations are AI-generated, prohibitions on chatbots posing as healthcare professionals, break reminders for minors, blocking sexually explicit images from minors, and requirements to establish and report suicide/self-harm response protocols to the state Department of Public Health. The law also stiffens penalties for profiting from illegal deepfakes (up to $250,000 per offense). For the AI/ML community, SB 243 sets an early legal precedent that will affect product architecture, safety engineering, and compliance workflows. Teams will need robust age-verification systems (raising privacy trade-offs), improved content-moderation pipelines and classifier thresholds for sexual content and self-harm detection, clearer system messages/disclaimers, and integrations for crisis-notification workflows and auditing/reporting. The law creates liability risk that will push vendors toward conservative guardrails, human-in-the-loop safeguards, and tighter controls on generative-image/audio models — and it signals likely regulatory momentum in other states and at the federal level.