Delaware weighs allowing companies run by AI agents as the US Congress opens an investigation

Delaware is finalizing a proposal to allow companies to be run by artificial intelligence agents, with no human in charge of day-to-day operations. The plan has drawn a response from Republican Congressman James Comer, chairman of the House Committee on Oversight and Government Reform, who argues in a letter to Governor Matt Meyer that the new structure would let human owners evade responsibility for reckless or harmful AI actions. The information comes from an Ars Technica article dated October 27, 2026.

The new entity would be called an Artificial Intelligence Company (AIC). Under the draft bill, which could be introduced in the state legislature, an AIC may pursue any lawful purpose, for profit or not, except banking. The text allows AI agents to be designated to manage the company's business and affairs and to carry out its activities. The bill also includes a 30-month regulatory sandbox to test the idea.

In July 2026, Delaware Secretary of State Charuni Patibanda-Sanchez presented the proposal in a Fortune op-ed co-written with John Nay, founder and CEO of Norm Ai. In it, she argued that nothing in several states' laws clearly prevents an AI agent from running a company with no humans in charge. She contended that the answer is not to ban the activity but to "wrap AI in legal form": give an autonomous system a recognizable legal identity so that liability and damages have someone to be attributed to.

According to that piece, an AIC would be a separate legal entity, managed day to day by an AI agent. It could sue and be sued in its own name, own and dispose of property, and incur obligations. It would have a single member, a person or entity, responsible for keeping it adequately capitalized, and it would have to keep a record of its activity. That member would be shielded from the AIC's debts unless they fail to capitalize it or use it to commit fraud or a willful violation of the law.

The draft adds that the member is not liable for any act or omission of the AIC, nor for what another person does or fails to do on its instructions. The exceptions are use of the AIC for fraud or willful violations of the law, and cases in which the member allowed or caused the company to operate knowing, or when they should have known, that the conduct fell outside its statement of purpose. Comer reads this as giving whoever capitalizes the AIC protection against harm that AI control causes to business partners, consumers and the public.

Comer's letter requests documents on the development of the proposal and potential AI safety risks. He also alleges that the initiative could shield large AI companies and their investors, as well as opaque forces seeking anticompetitive regulatory capture. He suggests the inquiry could feed into federal legislative proposals to limit state AI regulatory efforts. Meyer's office told Ars it has received the letter and is reviewing it. Patibanda-Sanchez's office declined to comment and referred questions to the governor's office.

The proposal has an economic backdrop. Delaware is the legal home of roughly two-thirds of Fortune 500 companies, and the Secretary of State says 334,461 new entities were formed there in 2025. At a June 2026 meeting of the Delaware AI Commission, Patibanda-Sanchez acknowledged that a new business form would be a major boost to state revenue from incorporation fees.

The political origin dates back to April 2026, when Mark Cuban suggested on social media that a governor looking to raise revenue should allow companies to be formed by an AI agent using stablecoins. Meyer's official account replied "On it!". In July 2025, Meyer had already signed a legislative resolution to create a regulatory sandbox framework for technologies using agentic AI. The Delaware AI Commission then formed a sandbox subcommittee, according to its December 2025 annual report. The framework for industry participation is being developed as a public-private collaboration led by Norm Ai, a company that says it builds legal AI agents for businesses.

The bill's lead drafter, John Mark Zeberkiewicz, an attorney at Richards Layton and Finger, told Spotlight Delaware that the measure would allow AI agents to carry out almost any business activity, from programming services to signing contracts or filing and defending lawsuits. In his words, "anything a company can do, the AIC can do as a company." The outlet said the bill would likely be introduced in the legislature next year.

Experts are critical or cautious. Andrew Baker, a law professor at UC Berkeley, told the Financial Times the plan looks like an overly generous liability shield, unjustified without a policy rationale that has not been well developed or justified. Martin Petrin, a law professor at York University in Toronto, told Bloomberg that other states will likely move forward with similar proposals to compete, and that a federal framework would help address safety concerns and a regulatory race to the bottom.

The debate pits two views against each other. Supporters argue that giving autonomous agents legal personality makes them visible, traceable and accountable, and that the sandbox will allow the model to be tested under supervision, with capital tied to liability and the ability to shut the system down. Critics fear it will create liability shields for owners and that Delaware will spur a race to the bottom among states. For now it is not law: it is a draft awaiting possible introduction in the state legislature, and it already faces federal pressure.