What changed
Delaware is considering a new corporate form, the Artificial Intelligence Company, that could let AI agents manage a business without a human at the helm. The proposed entity could operate for almost any lawful purpose except banking, with a 30-month regulatory sandbox, activity logs and a member generally shielded from the company’s debts unless it fails to provide adequate capital.
The plan now faces congressional scrutiny. Representative James Comer asked Delaware Governor Matt Meyer on October 26 for documents about the proposal, warning that it could let AI companies and their investors avoid responsibility for reckless harm. He also raised the possibility of federal legislation limiting states’ authority to regulate AI entities.
Why it matters
This is not just a new company registration option. It is a proposed answer to a hard legal question: if software signs contracts, pays suppliers, buys and sells property, or gets sued, who is actually responsible?
Delaware’s proposal would make the AI-run company a legal person in its own right. That could make autonomous systems easier to deal with in court and easier to trace through mandatory activity logs. It could also create a liability gap if the company is thinly capitalized and its human member remains protected from most debts.
That tension is now the central issue. Delaware sees a chance to attract more entity fees and tax revenue by giving autonomous software a formal legal home. Comer’s inquiry suggests Washington may see the same framework as a way for powerful AI firms and their backers to put distance between themselves and harmful decisions.
The practical outlook is unsettled. The draft may still change before reaching the legislature, and its final liability rules and safeguards remain unresolved. If the proposal advances, the key question will be whether the AIC’s legal identity makes responsibility clearer, or simply gives responsibility somewhere smaller to land.
Comments
No comments yet.