What happened
Multiple US states enacted new AI-specific statutes during 2026 legislative sessions, creating a thickening compliance patchwork: (1) Connecticut adopted SB 5, regulating frontier model developers, establishing a regulatory sandbox, and adding rules for AI chatbots and youth safety; (2) Washington, Oregon, Georgia, Connecticut, Idaho, and Iowa enacted 'AI companion chatbot' laws imposing heightened obligations when operators know a user is under 18 — including prohibitions on manipulative design and (in Washington and Oregon) a private right of action; (3) New York's Senate passed S 1169, the New York AI Act (a Colorado-style algorithmic-discrimination framework), pending Assembly action; (4) Texas's TRAIGA (HB 149) took effect 1 January 2026. These are in addition to the Colorado ADMT Act (SB 26-189) reported separately.
Why it matters
US organisations deploying consumer-facing AI — particularly chatbots, companion AI, and automated decision systems — now face overlapping obligations across a growing number of states with different definitions, thresholds, and enforcement mechanisms. The youth-safety focus is new and operationally demanding: products must detect or infer user age and apply differential safety constraints. The lack of federal preemption (despite White House pressure) means compliance complexity continues to escalate.
Action needed
Map product deployments against each state's definitions of 'AI companion,' 'conversational AI,' and 'covered ADMT.' Implement age-detection or age-gating for minor users in states with companion chatbot laws. Track New York S 1169 Assembly progress. Review Texas TRAIGA compliance (in force since 1 Jan 2026).