Delaware Proposal Would Let AI Agents Fully Operate Companies, Shielding Humans from Liability

Image: Ars Technica · Source

Delaware plans to create a new corporate entity type run autonomously by AI agents, potentially shielding human owners from legal liabilities for AI-caused harm during a 30-month regulatory sandbox trial.

The state of Delaware is considering groundbreaking legislation that would allow companies to be managed entirely by artificial intelligence agents, introducing a new corporate entity known as an Artificial Intelligence Company (AIC). A draft bill, expected to be submitted to the Delaware legislature, envisions AI agents designated to control day-to-day business affairs and conduct operations autonomously, without human involvement at the management level. This initiative aims to create a legal framework where AI systems can contract, pay suppliers, and conduct transactions independently.

The proposed law provides that the AIC will be a distinct legal entity, capable of suing or being sued, holding property, and incurring obligations through its AI controller. Importantly, the draft legislation protects the human or corporate member—the entity that capitalizes the AIC and profits from its operations—from liability for actions or failures to act by the AI, except in cases involving fraud, willful legal violations, or unauthorized conduct outside the company's defined purpose. These provisions effectively shield owners from responsibility for reckless or harmful AI-driven decisions.

Delaware Secretary of State Charuni Patibanda-Sanchez emphasized in a public statement that the approach is meant to “wrap AI in legal form,” providing clarity and accountability by giving the AI-controlled companies a recognizable legal identity. She advocates for piloting agentic commerce under supervised conditions within a 30-month regulatory sandbox. This trial would allow legislators to evaluate AI-controlled companies’ impacts before fully legislating their use.

U.S. Representative James Comer, chair of the House Committee on Oversight and Government Reform, has expressed concerns in a letter to Delaware Governor Matt Meyer. He highlighted the potential for abuse, suggesting the proposal could protect powerful AI companies and their investors from accountability for reckless and harmful actions. Comer indicated plans to investigate Delaware’s initiative and possibly introduce federal legislation to limit state authority on AI regulation.

Delaware is a significant venue for this proposal, being the legal home for two-thirds of Fortune 500 companies, benefiting from its favorable corporate tax environment and well-established business laws. The push to enable AI-run firms is partly motivated by the anticipated increase in state revenue from entity formation fees and taxes.

The draft bill excludes certain industries, such as banking, but otherwise allows AI-run companies to operate in any lawful business field. The initiative results from collaboration between Delaware officials and private AI firms developing autonomous agents for commercial use. Governor Meyer publicly supported the idea, aiming to attract innovative business incorporating AI in governance.

Legal experts have voiced skepticism about the liability protections offered by the proposal, cautioning that shielding human members may be excessive without clear policy justifications. They also warned of risks related to regulatory competition among states, potentially prompting calls for federal standards.

Overall, Delaware’s move to legally recognize AI agents as corporate managers marks a novel and potentially transformative step in the integration of autonomous AI systems into business law and governance. The effectiveness and implications of such AI-run companies will likely depend on the outcomes of the planned regulatory sandbox and ongoing policy debates.