Delaware drafts a company type AI agents could run, and Congress pushes back
Delaware is drafting a law for a new Artificial Intelligence Company (AIC) whose day-to-day business would be managed by an AI agent, tested first in a 30-month regulatory sandbox. Each AIC would have one member, a person or company, who funds it and is largely shielded from its debts and the agent's actions, with exceptions such as fraud, wilful law-breaking or failing to keep it adequately funded. AICs could do almost any lawful business except banking. US Rep. James Comer, who chairs the House Oversight Committee, has written to Governor Matt Meyer asking for documents, saying the plan could let owners escape responsibility for reckless AI. Nothing is law yet: a June report said the bill is likely to be introduced next year.
Why it matters: If you are building agents that sign contracts or pay suppliers, this is an early look at how the law may treat them. Who carries the liability when an agent acts is still unsettled, so keep a human accountable for now.
Ars Technica ↗AI-assisted, reviewed by Sorted AI
▲ 0
0 comments