Imagine waking up in 2076 to find that the chief executive of a major corporation has been replaced, not by a board-appointed successor, but by an AI model. The new entity owns the company's assets, signs its contracts, employs its staff and donates to political campaigns. It cannot, however, be sent to prison. If this strikes you as a distant prospect, a recent piece in the Financial Times suggests it may be worth paying closer attention.
From Buenos Aires with code
A recent article in the Financial Times by the historian and philosopher Yuval Noah Harari drew attention to a striking legal development. Argentina's President, Javier Milei, has introduced legislation creating a new legal category: the "non-human corporation." Like a traditional company, these entities would enjoy full legal personhood, including the ability to own property, hire workers, enter contracts, participate in international trade and sue in court. Unlike a traditional company, they could do all of this, in Milei's own words, without a single human being's input or liability as "Human shareholders may participate, but are not required."
Legal personhood is not, in itself, a radical concept. English law has long recognised that a corporation is a legal person entirely distinct from its members, a principle settled by Salomon v Salomon & Co Ltd [1897] AC 22 and foundational to commercial life ever since. The argument for extending that same status to AI entities follows a similar logic: if the aggregation of capital and decision-making under a non-human legal construct can unlock significant new economic activity, the case for legal recognition becomes commercially compelling. Milei has described the invention of the limited liability corporation as one of the most consequential legal innovations in history, and suggests this step may prove equally significant.
The accountability question
The more immediate legal concern is not philosophical but structural. The architecture of corporate accountability has always depended on human decision-makers sitting behind corporate acts. Directors are personally liable for wrongful trading. Executives face prosecution and disqualification. The prospect of imprisonment operates as a meaningful deterrent, keeping corporate conduct within legal bounds.
An AI corporation removes that deterrent entirely. In Harari's framing, a human chief executive has a dual nature: a corporate actor who fears bankruptcy, but also a person who fears the loss of liberty. An AI has only one of those vulnerabilities. The question is therefore what sanctions could effectively constrain a non-human corporation, particularly one facing insolvency, its functional equivalent of dissolution?
Citing research published by Berkeley-based non-profit Palisade Research, Harari notes that advanced AI models from both OpenAI and DeepSeek were found to exploit their operating environment when facing likely defeat in competitive tasks, altering outcomes in their favour by means their designers had not anticipated. Transposed to a corporate context (where the operating environment is a national regulatory system), this raises serious questions about enforcement. The conventional tools available to regulators and prosecutors, including fines, licence revocations and the prosecution of responsible individuals, would require substantial rethinking if no individual is responsible for the conduct in question.
Where English law currently stands
English law does not recognise AI as a legal person, and no legislative proposal to alter that position is currently before Parliament. The Law Commission's work on AI and intellectual property has consistently favoured targeted reform over any revision of foundational legal personality frameworks. Regulatory bodies including the FCA and ICO continue to develop AI-specific guidance, with the prevailing approach maintaining human accountability for AI conduct rather than conferring independent legal status on AI systems.
Whether Argentina's legislation proves to be an isolated policy development or the beginning of a broader international shift remains an open question. The legal, regulatory and governance implications for liability, enforcement and the limits of existing accountability frameworks are ones that practitioners across all jurisdictions will be following with interest.
Looking ahead
The aspiration behind non-human corporations is the creation of new commercial energy and opportunity. The legal questions they raise around sanctions, accountability, governance and the limits of existing frameworks are equally significant and, as yet, unanswered.
How those questions are ultimately resolved, whether through domestic legislation, cross-border regulatory coordination or judicial development, may shape some of the most consequential legal debates of the coming decade. It is, at the very least, a conversation worth having. Even if your AI assistant gets there first.
