US House Oversight Committee investigates Delaware's AI-run company plan

The US House Oversight and Government Reform Committee opened an inquiry into Delaware's plan for AI-run companies on October 26, 2026, when Chairman James Comer sent Delaware Governor Matt Meyer a letter raising liability-shield concerns.

Comer's committee is investigating the State of Delaware's efforts to establish a new form of corporate entity known as an Artificial Intelligence Company, under which an AI agent would run day-to-day operations of a corporate entity while providing the human or entity behind the AIC with liability protections from potential harms to consumers, the public, and business partners perpetuated by the controlling AI.

What it is

Delaware lawmakers could consider first-in-the-nation legislation next year that would create companies run by an Artificial Intelligence agent instead of a person.

Delaware Secretary of State Charuni Patibanda-Sanchez and Norm AI founder and CEO John Nay announced the proposal, along with their public private partnership, in Fortune magazine in July.

Comer's letter also names the bill's drafter, a lawyer practicing at Delaware based firm Richards, Layton, & Finger, whose biography describes him as having represented Facebook (now Meta) and who is described in the letter as the principal drafter of the proposed legislation.

What it requires

Comer's letter asks Meyer's office for all documents and communications about the AMC plan since Jan. 1, 2025, including internal AI chat logs.

Delaware's own draft would create a regulatory sandbox to test the AMC structure, including liability limits for parent companies and investors.

Comer wrote that his committee is concerned Delaware's AIC proposal could be used to protect powerful AI companies and their investors from responsibility for reckless harmful actions, while also protecting non-transparent forces seeking anticompetitive regulatory capture behind state AI initiatives, and said the inquiry could inform possible federal legislative efforts to limit such state AI regulatory initiatives.

What security leaders should do

Security and risk leaders should treat the AIC model as an early test of how liability and oversight get assigned when an autonomous agent, not a person, runs daily operations. Anyone evaluating a partner or vendor that might use such a structure should ask who holds legal and financial responsibility if the agent causes harm, and should require audit logs and human review regardless of how liability is allocated. Because this is a federal inquiry into a still-developing state proposal rather than a finished law, leaders should track Delaware's legislative process and any resulting congressional action before relying on an AIC-style structure for real operations.