Tag Archives: Business Ethics

Can a Company Be Run By AI Without Anyone Being Responsible?

Reading Time: 4 minutes

Recently Delaware proposed a new framework called “Artificial Intelligence Company” (AIC) that may change the way companies work. It was designed to let AI agents run daily business operations (sign contracts, file lawsuits, and make business decisions). It is not a law yet, but it raises a simple question: could a company exist as a legal person but without any person steering it?

What’s The Real Dilemma?

For a long time, corporations would work like this: investors get limited financial liability, but human leaders can be held accountable for all harmful decisions. However, Delaware’s proposed AIC framework turns it upside down. Under it, the company itself signs contracts and sues, but an AI agent decides what it does.

By letting an AI agent run a company while shielding the owner with limited liability, the proposal limits the responsibility of the people behind it. The protection has limits, but it is still a big change.

Who Pays If AI Messes Up?

The AIC itself could be sued, but an AI cannot go to prison or feel guilt. If a bot makes some damages and the company does not have funds to cover the damage, someone has to pay. In the proposal an owner must keep the company adequately funded, and is protected from its debts unless they fail to keep it adequately funded or use it for fraud or a willful violation of law.

Think of a case when an autonomous bot goes rogue, whether it makes illegal trades, breaches contracts or hacks competition’s infrastructure. That creates a chaotic situation in the legal sense. If the company itself lacks the funds to cover the damage, victims are left without any compensation and possibly without any human to hold accountable. This can happen even if the owner follows the rules, because legal scholars writing in Bloomberg Law say the proposal does not make clear how much money is enough.

Critics, including the House Oversight Committee, worry that the proposal could protect strong AI companies and their investors from responsibility for harmful actions. Law professor Andrew Baker, quoted in the Financial Times, said the most likely outcome appears to be an overly generous liability shield. My concern is that investors take the profit while society absorbs bad outcomes.

Even if the shield protects investors, there is a related problem for ordinary companies. An AIC has no board, but companies that already exist could give AI control too. In business law, company leaders have a duty to oversee what happens inside their business. They can’t just give control to an AI model and walk away. If leaders give a bot autonomy to make decisions on its own, shareholders could argue that they are failing in their job to control and oversee the business. In a 2026 working paper, Martin Petrin says directors’ and officers’ duties around AI are still unsettled, which is part of the problem.

People Should Be Accountable

If AI is a tool rather than an independent mind, how should the law handle it? Yale law professors Ian Ayres and Jack Balkin, in the University of Chicago Law Review Online, argue in their 2024 paper that the law should not depend on what an AI intended, because it does not have intentions. Instead, it should hold the people and organizations that build and use AI responsible. It makes sense for me, because if a company deploys a model without enough safety guards, proper testing, or human oversight, the blame should be theirs.

This keeps companies from hiding behind software and forces the people and organizations behind them to answer for what they create and use. In an AIC, the closest person is probably the owner, who funds the company. But the proposal protects the owner from debts.

Conclusion

Delaware’s proposed AIC framework tries to make it legal to run a company with AI. But the wish to automate companies cannot come at the cost of people’s safety, and legal accountability.

A company is not only to generate profit, it is tied to human responsibility and without it, it becomes a money machine possibly without anyone to blame for the disasters that may happen.

In my opinion you can let software do all the work, but humans must always provide oversight and take the blame if something bad happens.

AI Disclosure

Engine: Gemini 1.5 Pro
Used to help choose the topic and find the sources, gave feedback on the opening, and wrote draft sections which I used as a base and then rewrote in my own words. It also checked the full draft against the sources and proposed fixes.

  1. “I’d like to write a blogpost about the topic: ‘Can AI Run a Company Without Anyone Taking Responsibility’. I’m sending you a link to the source of the idea [link], and please find more sources that could allow me to write about this. I’d like to focus on corporate leadership, and who answers for harmful decisions, rather than general AI productivity.”
  2. “[text] Does it look right?”
  3. Use the Chicago Law Review perspective to argue that AI code is merely a tool of corporate intent, meaning liability must ultimately anchor to human capital and corporate treasuries.” (wording suggested by Gemini and pasted back as a prompt)
  4. “If AI is a tool rather than an independent mind, how should the law handle it? [text]”
  5. “what should be the next section”
  6. “write conclusion please”
  7. “[draft] Please check it against the sources and if there is something wrong, please propose fixes.”

Sources

  1. Charuni Patibanda-Sanchez and John Nay, “Exclusive: Delaware proposes testing the AIC, a new legal entity for agents in a regulatory sandbox,” Fortune, July 14, 2026.
  2. James Comer, Letter to Delaware Governor Matt Meyer, House Committee on Oversight and Government Reform, October 26, 2026.
  3. Joe Miller and Sujeet Indap, “Plan for AI ‘agents’ to run US companies challenged by lawmakers,” Financial Times, October 26, 2026.
  4. Ian Ayres and Jack M. Balkin, “The Law of AI is the Law of Risky Agents Without Intentions,” University of Chicago Law Review Online, November 27, 2024.
  5. Martin Petrin, “AI Accountability from the Outside In: Corporate Liability, AI Governance, and Managerial Duties,” ECGI Working Paper No. 935/2026, May 2026.
  6. Christopher M. Bruner, Sergio Alberto Gramitto Ricci, and Martin Petrin, “Delaware AI Company Proposal Needs Details Before Road Testing,” Bloomberg Law (opinion), August 3, 2026.

 

Tagged , , , , ,