House Oversight Chair Comer investigates Delaware plan for AI-run companies

House Oversight Committee Chairman James Comer, a Kentucky Republican, is investigating Delaware's proposed corporate structure for entities managed by artificial intelligence agents, the committee announced Monday. Comer sent Delaware Gov. Matt Meyer a letter expressing concern about the proposed Artificial Intelligence Company, or AIC. Comer warned that the structure could shield powerful AI companies and investors from responsibility for reckless, harmful conduct. Comer also warned that the proposal could protect opaque interests seeking anticompetitive regulatory capture through state AI initiatives.

Comer requested all documents and communications concerning development of the AIC since January 2025, with a November 9 deadline. The request includes internal AI chat logs, agreements with outside contributors, records concerning safety issues and every draft of the proposal. The committee said the material could help shape federal legislation, including measures to restrict state AI regulatory initiatives.

Comer's letter scrutinizes private-sector participation in developing the proposal. The letter says Delaware's AI Commission issued a December 2025 report without mentioning any work to evaluate safety risks or consult independent experts. The letter also says the Secretary of State's office declined to explain to reporters the office's working relationship with an AI company participating in the project.

Delaware Secretary of State Charuni Patibanda-Sanchez and Norm Ai CEO John Nay introduced the proposal through a Fortune opinion piece in July. Patibanda-Sanchez and Nay described a public-private partnership headed by Norm Ai, a legal AI company. An AIC would have the capacity to bring lawsuits, face lawsuits, own property and assume obligations under instructions from an AI agent.

Each AIC would have one member charged with maintaining adequate capitalization. That member would be protected from the AIC's debts unless the member failed to provide adequate capital or used the AIC for fraud or a deliberate legal violation. AICs would be permitted to operate only in a regulatory sandbox lasting 30 months, supervised by a committee comprising state officials and outside experts. AICs would have to identify their status to counterparties. AICs would not be permitted to engage in banking.

Patibanda-Sanchez and Nay argued that giving autonomous systems a recognized legal identity would make the systems' behavior easier to observe, trace and hold accountable. However, some legal scholars have challenged that argument. University of Georgia business law professor Christopher Bruner told WHYY in September that the proposal lacked an overarching regulatory framework for addressing consequences from AI-governed companies.

Patibanda-Sanchez rejected the characterization of the proposal as a liability shield. Patibanda-Sanchez told Bloomberg Law in July that the aim was to establish who bears liability when misconduct or psychopathic behavior occurs. Meyer also defended the initiative during a WHYY call-in program in September, arguing that the technology is coming and cannot be addressed by ignoring the technology. Draft legislation is expected next year. State lawmakers are expected to take up the draft when the lawmakers return in January.