Delaware Proposal for AI-Run Companies Draws Scrutiny From Federal Lawmakers
A Delaware plan to allow AI agents to run companies without human executives faces pushback from US Representative James Comer over legal liability and regulatory concerns.
Delaware is exploring a legal framework that would allow artificial intelligence agents to run entire businesses autonomously, triggering friction between state officials and federal lawmakers. Under a proposed draft bill, the state would create a legal entity known as an Artificial Intelligence Company (AIC). Within this structure, autonomous software agents could manage day-to-day operations, execute contracts, handle payments, and manage legal disputes on behalf of the business without a human executive at the helm. However, the initiative has drawn concern from the U.S. House Committee on Oversight and Government Reform, setting up a debate over state versus federal AI policy.
Federal opposition and liability concerns
In a formal letter sent to Delaware Governor Matt Meyer, Representative James Comer (R-Ky.), chairman of the House Oversight Committee, voiced strong objections to the proposal. Comer cautioned that the AIC structure could allow corporate founders and investors to evade legal responsibility for reckless or harmful decisions made by autonomous AI systems. The committee requested state documents concerning the creation of the framework and associated safety evaluations, hinting that federal lawmakers might consider legislation to limit state-level AI regulations.
Under the draft legislation, an AIC would feature a single "member"—an individual or corporate entity—who provides funding and retains profits. While the member must maintain adequate capitalization and log company activities, they are protected from debts and legal liabilities resulting from the AI agent's actions. Exceptions to this protection apply only if the member engages in fraud, commits a willful legal violation, or permits the AI to act outside its stated business purpose. UC Berkeley law professor Andrew Baker noted that the proposed liability shield appears overly generous without a clearly established policy rationale.
Sandbox framework and economic drivers
The initiative reflects a broader effort by Delaware leaders to capture emerging commercial activity driven by agentic software. In July 2025, Governor Meyer signed a legislative resolution establishing a regulatory sandbox framework to test AI applications across corporate governance, healthcare, and financial sectors. Later, in July 2026, Delaware Secretary of State Charuni Patibanda-Sanchez co-authored an op-ed with John Nay, founder and CEO of legal AI firm Norm Ai, arguing that giving autonomous systems a formal legal identity makes their actions traceable, visible, and legible to the law.
The proposed framework outlines a 30-month regulatory sandbox to evaluate agentic commerce under state supervision. Attorney John Mark Zeberkiewicz of Richards Layton and Finger, who served as a primary drafter of the measure, noted that AICs could carry out virtually any business task permitted for traditional companies, with the exception of banking. Proponents point to significant financial benefits for Delaware, which recorded 334,461 new entity formations in 2025 and already serves as the legal home for approximately two-thirds of Fortune 500 corporations.
What it means for developers
For software engineers and AI builders, Delaware's proposed framework signals a potential shift toward automated corporate operations where code executes real-world business activities directly. If enacted, developers building autonomous agent frameworks could deploy systems capable of signing contracts, ordering services, and holding assets without requiring human managerial approval for routine tasks.
As agentic workflows become more capable, cost-effective infrastructure remains an essential element for deployment. Developers interested in experimenting with autonomous software can try top AI models cheaply through one API at https://apixoai.online.
At the same time, the regulatory debate emphasizes key architectural requirements for future software agents. To comply with proposed oversight sandbox structures, developers will need to integrate strict activity logging, governance boundaries, and clear operational mandates into their software architectures. Martin Petrin, a law professor at York University, suggested that if other states adopt similar measures, federal rules may eventually be required to establish consistent baseline safety standards across the country.
Source: "No human at the helm": Delaware plan would let AI agents run entire businesses — Ars Technica AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
