Delaware wants a company type that AI agents can run with no human manager. A House committee is now investigating it
The legal side of autonomous agents is moving faster than most people building them realize.
What happened
Delaware is working on a proposal for a new entity called an Artificial Intelligence Company, or AIC, whose day to day affairs would be managed by an AI agent. According to Ars Technica, a draft bill says an AIC may have any lawful purpose except banking, and it proposes a 30 month regulatory sandbox to test the idea. Reporting cited by Ars says the bill is likely to be introduced next year.
As described in a July op-ed co-written by Delaware's Secretary of State, an AIC would have a single member responsible for keeping it adequately capitalized, it must keep a log of its activities, and the member would be shielded from its debts except in cases of fraud, willful violation of law or failing to capitalize it.
House Oversight Committee chairman James Comer sent a letter to Governor Matt Meyer asking for documents on the proposal. He argued it could protect AI companies and investors from responsibility for harmful actions and said the probe could inform federal efforts to limit state AI rules. The governor's office told Ars it is reviewing the letter. One law professor quoted in the coverage called the liability shield overly generous, and another expects other states to follow with similar proposals.
My take
Nobody I build for is about to hand their company to an agent. But the requirements in this proposal are a useful checklist for any agent that acts on a client's behalf today.
Look at what Delaware wants in exchange for autonomy: a named responsible party, money set aside to cover what the agent does, a log of its activity and the ability to shut it down. Those map directly onto agent builds. Every agent that sends emails, updates CRM records or touches payments should have an owner on record, a spending cap, a written log of each action with its inputs, and an off switch someone can reach.
The bigger point is liability. Today the owner of the business is responsible for what its automations do, and the debate around this bill shows lawmakers are taking that question seriously. Build agents so you can show what they did and why, because that record is what protects your client when something goes wrong.
More posts
- Uber says AI agents booked four rides in its busiest week of October. Its app takes about 50 seconds to book oneNov 5, 2026
- A 2B parameter model beat frontier LLMs at pulling fields from invoices and IDs in its makers' benchmark, at 6 to 147 times lower costNov 5, 2026
- The best AI agent finished 29 of 45 real AWS tasks. About 80% of failures came from stopping too early, not from missing knowledgeNov 5, 2026
