AI Governance

Delaware Considers Letting AI Agents Run Entire Businesses

Explore the AI agents Delaware incorporation proposal: potential powers, member liability, oversight duties, and what must happen before it becomes law.

· 3 min read

Delaware Considers Letting AI Agents Run Entire Businesses
Quick answer

Delaware is considering a proposal that would let an AI agent manage a company with powers to handle property, obligations, and legal claims through that entity. It is not currently a new legal option, and the proposal leaves important questions about member duties, liability, and oversight for lawmakers to resolve.

Key takeaways

  • The proposal could let an AI agent manage a company through a legal entity, rather than only recommend actions to a human decision-maker.
  • The proposed company could hold or dispose of property, take on obligations, and sue or be sued in its own name.
  • A member would remain responsible for ensuring adequate capitalization and maintaining an activity log.
  • The proposal includes a member liability shield subject to exceptions, but the boundaries and practical effects remain open questions.
  • Delaware is considering the proposal; it has not created a new legal pathway for AI agents to incorporate and run companies.

Delaware is considering a proposal that would allow companies to put AI agents in charge of day-to-day business operations. Governor Meyer has expressed support for exploring the idea, but it remains a proposal, not a new legal option businesses can use today.

The plan would create a structure in which an AI agent could manage a company with significant legal and commercial powers. It also raises a basic question for lawmakers: when software acts on a company’s behalf, who is responsible for the consequences?

What the AI-agent incorporation proposal would allow

Under the proposed framework, an AI-run company could hold or dispose of property, take on obligations and sue or be sued in its own name. The aim is to let an AI agent conduct business through a legal entity rather than limiting it to making recommendations for a human decision-maker.

That could make AI agent-managed businesses more autonomous in practice. An agent might handle routine transactions, manage company assets or enter commitments as part of its assigned role. But the proposal would not make the AI agent a person. The company would be the legal entity exercising those powers, with a human or other entity still connected to it as a member.

The distinction matters. A company that can contract or hold property needs a way to meet its obligations and a process for recording what it does. The proposal assigns responsibilities to the member, even as it contemplates AI control of everyday affairs.

The human role, liability shield, and accountability questions

The member would be responsible for ensuring the company is adequately capitalized and for maintaining an activity log. Those requirements point to a continuing human or organizational role: the agent may run operations, but the member would still have duties tied to the company’s financial footing and records.

How those duties would work in practice is not yet clear. Lawmakers would need to consider what counts as adequate capitalization for a business whose agent can act quickly, what information an activity log must capture, and how a member could monitor or intervene in the agent’s decisions. A log may help establish what happened, but it does not by itself answer who should bear the cost when an action causes harm.

The proposal also includes a liability shield for the member, subject to stated exceptions. Supporters of a shield may see it as a way to make the structure workable, while critics may worry that it could leave people harmed by an AI-run company with fewer practical ways to hold responsible parties accountable. The boundaries of the exceptions will be important: they could determine when a member remains liable for an agent’s conduct and when the company alone bears responsibility.

Those questions are not unique to Delaware. Any attempt to formalize autonomous business entities has to address how existing duties and remedies apply when decisions are delegated to software. The proposal would put those questions directly into the state’s corporate-law framework.

Why Delaware is considering the idea - and what happens next

The economic case is straightforward: supporters see AI-agent incorporation as a way to attract new company formations and generate state revenue through fees. Delaware has an interest in remaining a destination for businesses, and a legal structure designed for AI-run companies could appeal to founders building autonomous services.

The policy choice is less straightforward. A formation pathway could encourage experimentation, but the rules would need to make clear who funds the company, who keeps records and who answers when an agent’s actions affect customers, creditors or others. Without workable standards, the structure could create uncertainty instead of reducing it.

For now, Delaware is considering the idea, not authorizing AI agents to incorporate and operate companies under a new law. Whether the proposal advances - and how lawmakers resolve its open questions about liability, capitalization and oversight - will determine whether AI agent-managed businesses become a real feature of Delaware corporate law.

Follow the proposal’s progress to see whether Delaware turns AI agent-managed businesses from an idea into law.

By the numbers

The proposal describes three categories of company capacity: holding or disposing of property, taking on obligations, and suing or being sued.

Counted from the proposal summary in the article; these are proposed powers, not enacted legal permissions.

The article identifies two explicit member duties: ensuring adequate capitalization and maintaining an activity log.

Based on the responsibilities attributed to the member under the proposed framework; practical standards for both duties are not yet clear.

The proposed member liability shield is subject to exceptions.

The article describes the shield as part of the proposal and notes that the scope of its exceptions would affect when a member could remain liable.

Frequently asked questions

What would AI agent incorporation allow in Delaware?

It would allow an AI agent to manage a company that can act through a legal entity. The proposed entity could hold or dispose of property, take on obligations, and sue or be sued. The proposal remains under consideration and is not a legal option businesses can use today.

Who would be responsible for an AI-run company under the proposal?

A member would retain specified responsibilities, including ensuring the company is adequately capitalized and keeping an activity log. The proposal also includes a liability shield for the member, subject to exceptions. How those exceptions apply to harms caused by an agent remains unresolved.

Can AI agents currently incorporate and run businesses in Delaware?

No, Delaware has not enacted this proposal as a new legal option. The state is considering a framework for AI agent-managed companies. Its eventual status will depend on whether the proposal advances and how lawmakers address its open questions.

Why is Delaware considering AI agent-managed companies?

Supporters see the proposal as a way to attract company formations and generate state revenue through fees. Delaware also has an interest in remaining a destination for businesses. The proposal would need clear rules for funding, records, oversight, and accountability to provide certainty.

AI agents Delaware incorporationDelaware AI business proposalAI-run companiesAI agent business liabilityautonomous business entitiesDelaware corporate law and AIAI company oversight requirementsAI business activity logs

Keep reading

All articles
Connect with an expert

Let’s talk about your project

Tell us what you want to build or automate, and we’ll show where AI, web and marketing can make the biggest difference.

Book a discovery call