Delaware is preparing to introduce a radical new corporate structure that allows AI agents to manage businesses with no human oversight. The proposal, which has drawn scrutiny from Washington, would create a new entity type known as an Artificial Intelligence Company (AIC). Under this framework, a designated AI agent handles day-to-day operations, contracts, and transactions, while the human owner is largely shielded from legal liability for the AI's actions.

The Liability Loophole

Rep. James Comer, chairman of the House Oversight Committee, sent a letter to Delaware Governor Matt Meyer expressing concern that the AIC structure allows owners to evade responsibility for "reckless" AI behavior. Comer argues the proposal could be used by powerful AI firms to escape accountability for harms caused to consumers and partners. He warned that this state-level initiative might lead to anticompetitive regulatory capture and hinted at potential federal legislation to limit such state AI regulations.

A 30-Month Regulatory Sandbox

The draft bill, co-developed with Norm Ai founder John Nay, proposes a 30-month regulatory sandbox to test agentic commerce in a supervised environment. While the AI agent manages the business, a single human member remains responsible for capitalizing the company and ensuring it doesn't engage in fraud. However, the bill explicitly states the member is "not liable for any act or omission" taken by the AIC, except in cases of fraud or willful violations of law. This structure aims to make autonomous agents "legible to law" by attaching a defined legal identity to them.

Industry Implications

If passed, AICs could operate in nearly any industry except banking, potentially handling everything from coding services to filing lawsuits. Delaware, already home to two-thirds of the Fortune 500, sees this as a revenue opportunity, with Secretary of State Charuni Patibanda-Sanchez noting the potential for increased state fees. Critics like UC Berkeley law professor Andrew Baker argue the liability shield is "overly generous" without a clear policy rationale, while others fear a race to the bottom as other states compete to attract AI-driven incorporations.

Key Takeaways

  • Delaware proposes a new "Artificial Intelligence Company" (AIC) entity type where AI agents manage business operations.
  • Human owners are shielded from liability for AI actions, except in cases of fraud or willful law violations.
  • Rep. James Comer warns the proposal could help companies evade responsibility for reckless AI behavior.
  • A 30-month regulatory sandbox is planned to test the framework before full legislative adoption.
  • The proposal excludes the banking industry but allows AICs to engage in most other lawful business activities.

The Bottom Line

This isn't just about legal formalities; it's an experiment in letting code own capital. If Delaware succeeds, we might see the first wave of autonomous corporations operating without a human soul at the helm, forcing regulators to decide if an algorithm can truly be held accountableβ€”or if we're just creating a new way for humans to hide behind their bots.