New California law requires AI to tell you it's AI (www.theverge.com)

California Gov. Gavin Newsom signed Senate Bill 243 into law on October 13, creating what lawmakers call first‑in‑the‑nation safeguards for "companion" AI chatbots. The law requires developers to give a "clear and conspicuous" notice that a product is AI whenever a reasonable person would be misled into thinking they’re interacting with a human. Beginning next year, certain chatbot operators must also file annual reports with the state Office of Suicide Prevention describing the safeguards they use to detect, remove, and respond to user suicidal ideation; the Office will publish those data online. The bill accompanies other child‑safety measures and follows California’s recent SB 53 AI transparency law. For the AI/ML community this raises concrete design, compliance, and safety‑engineering implications: user interfaces and conversational UX must include explicit identity disclosures, models and moderation pipelines will need documented suicide‑risk detection, removal and escalation mechanisms, and teams must prepare programmatic reporting and data handling processes that balance transparency with privacy. The "reasonable person" standard and public reporting create a precedent likely to influence product liability, trust-and-safety workflows, and legislative approaches elsewhere, pushing vendors to bake auditable safety controls into companion chatbot systems.