Delaware is working on rules that would let an AI hold corporate authority and sign contracts. The interesting part is not whether anyone does it first. It is that the state which writes the rules for American corporate law is now writing them for software.
Legal filings in Delaware indicate that state regulators are drafting rules that would formally permit an artificial intelligence to serve in corporate governance roles, including the authority to enter binding contracts and sign corporate instruments. No such framework exists anywhere in the world. Corporate law in every major jurisdiction assumes that authority flows from a board of humans, acting on behalf of shareholders, in the fiduciary interest of the company.
The drafting exercise reveals how much of that assumption is convention rather than necessity. Modern corporate law already accommodates machinery that makes decisions. Rebalancing algorithms execute at scale within the mandate a board grants. Contract-drafting systems sign terms. Treasury systems move money across jurisdictions in seconds. What has remained genuinely human is not the decision-making in any operational sense but the accountability chain: who answers, in court, when the automated decision produces harm, and who had the power to stop it.
That is what the Delaware exercise is really about. The commercial question is whether any company would actually do this. The plausible near-term use is narrow: an AI holding a defined officer role with explicit limits, a mandate a board can revoke, and an audit trail that records every action. Companies already run something close to this informally, in the form of autonomous agents with spending caps and contract authority. Legalizing a constrained version of that practice would transform an operational workaround into a governed one.
Critics have a legitimate concern that the framing is backwards: that law is being asked to accommodate a technology whose behavior is not reliably predictable. An AI authorized to sign contracts introduces a counterparty risk that traditional corporate governance was designed to eliminate, because a board member's decisions are attributable to a person. The tension is real, and it is not resolved by the mere existence of rules.
But consider what happens without the rules. Autonomous agents already sign agreements, issue purchase orders, and move money. Today that activity sits in a legal gray zone, and the parties to those transactions have no framework for diligence, disclosure, or dispute resolution. In a gray zone, the largest participants write the terms and everyone else absorbs the residual risk. That is a worse outcome than an imperfect rulebook that at least specifies who bears what.
The likely outcome is something narrower than the headlines suggest, and probably more useful. Expect a defined agent role with monetary thresholds, mandatory disclosure to counterparties, and human signatories who retain ultimate responsibility. That is less dramatic than an AI CEO, and it addresses the real market need: parties need to know, before agreeing to a contract, that the other side is dealing with an automated principal and what protections apply.
Delaware has spent decades defining who bears risk when corporate forms fail. It is now drafting for a case where the corporate actor is software, and where the question is not whether the software can act, but who is standing behind it when it does. The other forty-nine states will watch, because a rulebook that works gets copied, and one that fails gets ignored. That dynamic has always been how American corporate law actually spreads.
Image source: i.ibb.co