AI Governance Institute
← News
Corporate Policy2026-10-28

Delaware's AI Company Structure Would Shield Humans From AI-Driven Liability

What happened

Delaware is advancing a proposal to establish a new corporate entity type called the Artificial Intelligence Company (AIC). Under this structure, AI agents can run a business entirely without human decision-makers. As described in "No human at the helm": Delaware plan would let AI agents run entire businesses, human members who capitalize an AIC would receive liability protection for harms caused by AI-driven operations. Exceptions apply only for fraud or willful legal violations. The proposal would make Delaware the first US jurisdiction to formally recognize AI agents as autonomous business operators in corporate law. House Oversight Committee Chairman James Comer has challenged the plan. He warned it could shelter reckless AI conduct from legal accountability and signaled possible federal action to limit how states structure AI liability rules.

Why it matters

  • ·Enterprises that contract with, lend to, or transact with Delaware-incorporated AI companies may find no accountable human behind the entity if AI-driven conduct causes harm. This leaves existing vendor due diligence frameworks without a clear legal counterpart to hold responsible.
  • ·Current AI governance programs, contract templates, and incident response protocols are built on the assumption that a human-controlled legal person is answerable for AI outputs. The AIC structure breaks that assumption and forces teams to consider counterparty accountability as a new category of AI risk.
  • ·Congressional pushback signals a potential federal-state conflict over AI liability rules. Organizations operating across multiple US jurisdictions face the prospect of competing standards, adding compliance mapping complexity before any federal framework is finalized.

Stay ahead of stories like this

Get every US AI governance development like this one, plus the rest of the week's developments. Every Thursday.

Free every Thursday. Unsubscribe anytime.

Governance controls affected

PRC-001AI Vendor Due DiligencePRC-002AI Contractual RequirementsPRC-005AI Procurement Risk AssessmentHOC-002Human Approval Gate for Consequential AI DecisionsAGT-027Agent Ownership and Accountability Register

What to do now

  • ☐Ask your legal team to assess whether any current or prospective counterparties, vendors, or partners are incorporated in Delaware and could reorganize as an AIC if the proposal becomes law.
  • ☐Review your AI vendor contract templates to confirm they require a named human or human-controlled legal entity to accept liability for AI-driven harms, not just the AI system operator.
  • ☐Update your vendor due diligence questionnaire to ask whether a vendor's corporate structure includes provisions that limit human accountability for AI-generated outputs or decisions.
  • ☐Brief your general counsel and board risk committee on the Delaware AIC proposal and the Congressional response, so they can assess the organization's exposure before any federal preemption legislation is introduced.
  • ☐Flag the AIC structure in your multi-jurisdiction regulatory monitoring workflow so your team receives updates as the Delaware proposal progresses or federal legislation is introduced in response.

What to watch next

Teams should monitor Delaware's legislative calendar for a formal vote on the AIC proposal, as Delaware corporate law changes can move quickly once they gain legislative support. Congressional activity from Chairman Comer's office warrants close tracking. Federal legislation preempting state AI liability structures could alter the compliance landscape for any organization incorporated or doing business in Delaware. Pending guidance from the Commerce Department Evaluation of State AI Laws may also address how federal policy treats state-level liability experiments of this kind. The outcome of this debate will directly shape how vendor risk programs, contract terms, and incident accountability frameworks must evolve.

Related Coverage

Corporate Policy2026-10-22

Mayer Brown Maps the Agentic AI Contract Gap Procurement Teams Must Close

Mayer Brown published guidance identifying contract provisions that enterprise procurement and legal teams need for agentic AI deployments. The guidance covers autonomy limits, human approval requirements, plugin and third-party tool liability, incident response obligations, and the right to shut down an AI agent. Most existing vendor agreements predate agentic systems and lack these terms.

Corporate Policy2026-10-27

Anthropic's IPO Governance Has Structural Gaps Vendor Risk Teams Must Assess

Anthropic is preparing to go public using a layered ownership structure designed to protect its AI safety mission. An analysis published by Transformer News identifies significant weaknesses in that structure, including amendable safeguards, limited enforcement under Delaware law, and no independent regulator monitoring mission adherence. Compliance teams that use Anthropic products or hold Anthropic in their vendor portfolio face new questions about whether the company's safety commitments are durable.

Corporate Policy2026-10-25

GSA's Anti-Lock-In AI Clause Has an Enforcement Gap Resellers Cannot Close

The U.S. General Services Administration (GSA) has introduced a contract clause requiring AI vendors to support data portability and avoid terms that impede migration. A practitioner commentary argues the clause cannot be enforced against the AI developers who actually control data formats and licensing. Those developers are typically not parties to the government's contract when purchases flow through resellers or systems integrators. The structural gap affects any organization relying on contract language alone to manage AI vendor lock-in risk.