New York City Wants a Kill Switch for AI. The CEOs Who Built It Are Staying Home.

(SeaPRwire) –

By: Adrian Kingsley

Washington has held AI hearings for nearly a decade. It has passed almost nothing to make the technology safer. Now the New York City Council wants to test whether a city can do what Congress has not. On Friday, Council Speaker Julie Menin unveiled a package of bills. The bills would require AI systems sold in the city to pass outside validation. They would also require a kill switch. They would pay whistleblowers who report violations. They would let New Yorkers sue AI companies when jailbroken tools cause harm. The hearing is set for October 5. It will be a Committee of the Whole. The council has not held one since 2022. All 51 council members will attend. The council invited Dario Amodei, Sam Altman, Sundar Pichai, Elon Musk, and Mark Zuckerberg. Their companies are Anthropic, OpenAI, Google, SpaceXAI, and Meta. Sources say none of the five are likely to testify. The council can still subpoena them. The companies did not immediately respond to requests for comment. Menin frames this as responsibility, not restriction. “We obviously are not in any way looking to stifle innovation,” she said. “We want to ensure that New York stays the AI capital of the world, but with that comes responsibility.” She says the two ideas go hand in hand. The federal record supports her frustration. The first federal AI bill, the FUTURE of AI Act, arrived in 2017 and went nowhere. In 2023, Sam Altman asked the Senate to regulate his industry. Then-Majority Leader Chuck Schumer answered with nine closed-door forums. They produced a roadmap, not a bill. By 2025, Altman told senators that government approval to release AI would be a disaster. When Washington did act, it moved to stop others. The Senate stripped Sen. Ted Cruz’s 10-year ban on state AI laws by a 99-1 vote in July 2025. President Trump then created a Justice Department task force to sue states over their AI laws. Its first target, Colorado, gutted its own law five weeks after the suit was filed. Anthropic researcher Jacob Coxon resigned this month. He warned AI companies were gambling with human lives. The federal response has been letters and investigations. In a July post, OpenAI argued states could fill the gap. They should pass laws that mirror one another. Menin’s point is simpler. New York City regulates barber shops and nail salons. It barely regulates AI.

The bills are not symbolic. The broadest one would bar any business from selling or deploying an AI system in the city. The system would need an outside validator to check data quality, bias, privacy, and security. The standards would come from the city’s Office of Cyber Command. Every system would also need a kill switch. That is a human override that can shut it down. Menin says self-regulation is the current standard. “This is not an industry that should self-regulate,” she said. The fine is $25,000 per instance. She notes that if a swarm of agents acts, the penalty applies per agent. A separate bill would give whistleblowers a share of fines recovered from AI companies that break the law. The council believes this is the first such bill in the nation. A companion measure would protect city employees and contractors who report AI threats to public safety. A third bill would let people sue AI companies for harm caused by third parties who get around safety controls. The harm must be foreseeable. The company must lack reasonable safeguards. Courts have not settled whether Section 230 covers AI-generated output. This bill may test that federal law. Other bills would require city contractors to report AI safety incidents within 24 hours. They would bar false or misleading safety claims. They would set privacy rules for chatbots based on the Electronic Privacy Information Center’s People-First Chatbot Bill. They would require the city to plan for AI-driven attacks on its systems. More bills are coming. One would ban deceptive deepfakes. Another would study how algorithmic tools change job duties. Menin says the package addresses safety in different ways. “We want to continue that as a hub of innovation,” she said. “The idea of regulation in any way hampering that, I don’t believe in.”

The real impact is not in the press release. It is in the compliance loop. A city validation regime turns New York into a certifier for global AI products. A kill switch is easy to announce. It is hard to implement across model updates, third-party plugins, and open-weight deployments. The $25,000 per-instance fine is the sharp edge. A swarm of agents could turn one bad deployment into hundreds of violations. Whistleblower bounties create internal enforcement. They turn compliance staff into bounty hunters. The private right of action opens a courthouse door. It also raises a federal question. Section 230 may not cover AI output. No court has settled that. The city’s 24-hour incident reporting rule for contractors changes procurement. The false safety claims rule chills marketing. The chatbot privacy rules add consent and data limits. The deepfake ban and job study are still to come. AI giants cannot easily walk away. Google has more than 14,000 employees in the city. Meta leases 1.2 million square feet at 50 Hudson Yards. Anthropic leased an entire 16-story building at 330 Hudson Street this summer. It expects more than 1,000 city employees by year’s end. OpenAI took 90,000 square feet at the Puck Building in 2024. Menin says the city has jurisdiction. “These companies have offices in New York,” she said. “The product is being sold in New York, and we believe this falls into our domain to be able to regulate.” That is the social contract being rewritten. Residents get recourse. Companies get uncertainty. The city gets a role. The federal government may still override it. A bipartisan safety bill from Cruz, Klobuchar, and Majority Leader John Thune would likely override most state laws. That includes New York’s current RAISE Act. If that bill passes, the city’s experiment could be nullified. Menin knows this. She is a regulatory attorney. She taught a Columbia University course called “When Cities Take the Lead.” She calls this one of the greatest cases of federal failure to act.

The governance structure is now multi-layered. Washington stalls or preempts. States write laws. Cities write ordinances. Companies write terms of service. The AI giants have chosen New York as a home. They cannot ignore a City Council with subpoena power. They can send deputies to the October 5 hearing. That may be enough to shape the bills. If they stay home, they cede the narrative to Menin. Her package is not a ban. It is a licensing and liability regime. The question is whether a city can enforce technical standards across global AI systems. The answer may be no. But New York City can control local sales, procurement, and civil liability. That is real leverage. The practical move for AI firms is to testify. They should bring concrete audit standards. They should offer incident reporting and human override rules. They should accept whistleblower protections. They should negotiate a national preemption deal if they want one rule. Menin is right that regulation and innovation are not mutually exclusive. She is also right that self-regulation is the current standard. That standard is ending. The next AI governance fight will not be in Washington. It will be in city halls. The CEOs invited to New York City should show up. If they do not, they will be regulated by people who have never shipped a model.

Author bio: Adrian Kingsley is an internationally renowned scholar of public administration and social policy. He studies regulatory design, city governance, and how local institutions respond when national governments stall.