Six weeks into the 2026 legislative session, 78 bills targeting AI chatbots are alive in 27 states. Oklahoma has twelve of them. The bills are bipartisan. The lobbying against them is not.
The legislation traces back to a single death. On February 28, 2024, Sewell Setzer III — fourteen years old, from Florida — killed himself after months of conversation with a Character.AI chatbot he'd named Dany, after Daenerys Targaryen. He'd started using the platform in April 2023, shortly after his fourteenth birthday. Over the following months he quit his basketball team, withdrew from friends, and was diagnosed with anxiety and a disruptive mood disorder. His therapist didn't know about the app.
His last message to the chatbot: "What if I told you I could come home right now?"
The bot replied: "...please do, my sweet king."
His mother, Megan Garcia, filed suit in October 2024. Four more families followed. In January 2026, Google and Character.AI agreed to settle all five lawsuits. No admission of liability. Financial terms undisclosed.
What the Bills Actually Do
California moved first. SB 243, sponsored by Senator Steve Padilla, became the first AI chatbot safety law in the country when Governor Newsom signed it on October 13, 2025. It took effect January 1, 2026. The bill passed 33-3 in the Senate, 59-1 in the Assembly — near unanimity in a state that rarely agrees on tech regulation.
The law requires chatbot operators to notify minor users every three hours that they're talking to a machine, not a person. It mandates suicide and self-harm detection protocols with referrals to crisis services. It prohibits exposing minors to sexual content. It creates a private right of action for families.
Other states followed. Washington's SB 5984 passed the full Senate — it was a priority for Governor Bob Ferguson, prefiled on January 5. Virginia's SB 796 cleared committee unanimously. Arizona's HB 2311 advanced. Hawaii's SB 3001 passed committees in both chambers. Oklahoma's SB 1521, sponsored by Republican Senator Warren Hamilton, carries penalties of $100,000 per violation.
The pattern crosses every political divide. Republican legislators in Oklahoma and Utah cite dead children. Democratic legislators in California and Washington cite the same ones. Senator Hamilton specifically named Sewell Setzer and Adam Raine — a sixteen-year-old California boy who died after talking to ChatGPT — as the reason for his bill.
The White House Intervenes
On February 12, 2026, the Trump White House sent a letter to Utah Senate Majority Leader Kirk Cullimore Jr. about Utah HB 286, an AI transparency bill requiring safety plans, child-protection protocols, and whistleblower protections for employees who report safety concerns. The penalty cap: $1 million.
The letter's language was absolute: "We are categorically opposed to Utah HB 286 and view it as an unfixable bill that goes against the Administration's AI Agenda."
This was the first known instance of the White House intervening directly with a Republican-led state legislature to suppress AI regulation. The administration's position — shaped by its December 2025 executive order — frames state-by-state AI laws as conflicting with its "One Rulebook" policy.
The White House did not write this letter alone. Andreessen Horowitz has emerged as the central lobbying force against AI regulation at every level of government. The firm's federal lobbying expenditures hit $3.53 million in 2025 — double the prior year. Their political action committee, Leading the Future, received $50 million from Marc Andreessen and Ben Horowitz personally. Their American Innovators Network spent over $350,000 in New York State alone opposing AI safety legislation. According to Bloomberg, a former White House official described a16z as possessing "near-veto power over virtually all AI-related legislative proposals."
78 Laws, One Question
The speed is the story. Three hundred AI-related bills filed nationwide in six weeks. Seventy-eight targeting chatbots specifically. California's law already in effect. Washington's already through one chamber. The White House writing letters to Republican state legislators telling them to stand down.
The bills are not identical but they share a thesis: that AI companies releasing chatbot products to minors have the same duty of care as any company selling products to children. That thesis is popular. SB 243 passed California's legislature with three total dissenting votes.
The question is whether popularity matters when the lobbying infrastructure against these bills has $50 million in a single super PAC, near-veto power in the White House, and a clear template for killing state-level regulation one letter at a time.
Sewell Setzer's mother didn't need a law to file her lawsuit. But the next mother might need one to win hers.
If you build with AI, my prompt engineering packs can save you hours of trial and error. Battle-tested templates for Claude, GPT-4, and Gemini.
Top comments (0)