House Oversight Committee Chairman James Comer (R-KY) has launched an investigation into Delaware’s proposal to establish a new corporate entity type known as the Artificial Intelligence Company (AIC), which would be operated day-to-day by AI agents. In a letter sent to Delaware Governor Matt Meyer, Comer expressed concern that the plan could shield AI firms and their investors from liability for harms caused by autonomous systems. The committee has requested all documents and communications related to the AIC’s development since January 2025, including internal AI chat logs, agreements with outside contributors, and safety records, with a deadline of November 9.
The proposed AIC structure was unveiled in July by Delaware Secretary of State Charuni Patibanda-Sanchez and John Nay, CEO of legal AI firm Norm Ai. Under the plan, an AIC would possess legal personhood, allowing it to sue, be sued, hold property, and incur obligations under the direction of its AI agent. The entity would operate within a 30-month regulatory sandbox overseen by state officials and experts, requiring disclosure of its status to counterparties while barring participation in banking activities. A single member would be responsible for capitalization but would generally be protected from debts unless they failed to fund the company or engaged in fraud.
Comer’s inquiry highlights potential risks regarding "anticompetitive regulatory capture" and notes that a December 2025 report by Delaware's AI Commission did not mention independent safety assessments. While Delaware officials argue the framework aims to make AI conduct "visible, traceable, and accountable," critics like University of Georgia law professor Christopher Bruner have questioned the lack of overarching regulatory plans for fallout. Governor Meyer defended the initiative as necessary engagement with emerging technology, with draft legislation expected for consideration by state lawmakers in January.
This investigation signals growing federal scrutiny of state-level attempts to create bespoke legal frameworks for autonomous entities. By targeting Delaware, a jurisdiction central to U.S. corporate law, House Oversight Chair James Comer is challenging the premise that states can independently define liability structures for AI without broader national standards. The request for internal chat logs and private-sector contributor agreements suggests concerns about transparency and potential conflicts of interest in how these novel legal definitions are being drafted.
The core tension lies between innovation and accountability. Delaware’s proposal seeks to integrate AI into existing corporate forms through a supervised sandbox, aiming to clarify rather than obscure responsibility. However, the explicit shielding of owners from certain liabilities raises significant compliance and operational risk questions. If federal regulators view this as a mechanism for evading traditional corporate governance duties, it could lead to preemptive federal legislation limiting state authority over AI-specific corporate charters, potentially stalling the very innovation the sandbox intends to foster.
