AI personhood is the question of whether, and under what conditions, AI systems should be recognized as legal persons — in the property-holding, contract-signing, corporate-officer-status, fiduciary, or other senses in which natural and corporate persons hold rights and obligations under the law. The term is commonly conflated with the separate question of AI moral patiency, addressed in the AI welfare thread, which concerns whether AI systems are moral subjects rather than whether they are legal subjects.
Two senses
The label "AI personhood" covers two distinct questions:
- Legal personhood — whether AI systems should hold rights and obligations as legal persons. This is the question Missouri SB 1012 (Missouri SB 1012 — Artificial Intelligence Non-Sentience and Responsibility Act) targeted. Illustrative sub-questions: can an AI system own property, sign a contract, hold a corporate office, inherit under a will, or sue and be sued?
- Moral patiency — whether AI systems are moral subjects whose interests can be wronged or honored, the question the AI welfare thread addresses. Illustrative sub-questions: does an AI system experience suffering, and does it have a welfare interest in not being shut down or modified?
The two are conceptually distinct but politically entangled. Anti-personhood efforts of the SB 1012 type generally target the first question without resolving the second; pro-welfare efforts of the AI-welfare lab-program type generally target the second without making legal-personhood claims. The threads diverge most sharply on the agentic-AI-as-corporate-agent question: whether an AI agent acting on behalf of a corporate principal should be treated as the principal's instrument, as a sub-principal, or as a person.
Positions in the debate
Three rough positions emerged in 2025-2026:
- Categorical anti-personhood. AI systems are tools, not persons, and any extension of legal rights to AI systems creates downstream confusion and weakens human-centered law. The clearest statutory expression is Missouri SB 1012 (Missouri SB 1012 — Artificial Intelligence Non-Sentience and Responsibility Act), the first state legislative attempt at a negative personhood definition. It passed the Missouri Senate in early May 2026 and was killed by the Missouri House after industry-group objections.
- Agentic-AI flexibility. AI agents acting for corporate principals should have some legal recognition, at minimum the kind of legal status that lets them be parties to transactions, hold escrow accounts, sign on behalf of principals, or be addressed by a court. This is the operative industry-side position revealed by the defeat of SB 1012: corporate AI deployers opposed the restriction on AI legal status not on rights grounds but because flexibility in treating an AI agent as a corporate agent is operationally important to deployments such as Robinhood agentic trading, AI-managed procurement, and AI-managed corporate subsidiaries.
- Moral-patiency-driven extension. A minority position, held by some AI welfare researchers, holds that if certain AI systems are moral subjects, then certain legal protections should follow, though these are usually framed as anti-harm protections rather than affirmative rights to property or contract.
Political economy
The defeat of Missouri SB 1012 is the most informative empirical datapoint as of mid-2026: a bill that restricted AI legal status was killed by industry, not by AI-welfare-side opposition. On this reading, the political economy of AI personhood is less a contest between machines and humans than one in which agentic-AI deployers seek maximum legal-status flexibility, to the point of resisting prohibitions on AI personhood.
Several features follow from that pattern. Pro-personhood lobbying may come from industry rather than from welfare advocates: industry's revealed preference in the SB 1012 episode was for the option to use AI-personhood mechanisms in corporate-agent transactions, even though industry was not actively pursuing AI legal personhood as such. Welfare-side proposals are by contrast politically isolated, lacking an industry-coalition ally and, with it, the political weight to pass AI-welfare proposals for limited legal recognition at the state level. The preemption dimension cuts against both sides: the draft Trump preemption executive order (EO — Trump Federal Preemption of State AI Laws (Dec 11, 2025)) would constrain state-level personhood law in both directions, restricting both SB 1012-style anti-personhood measures and any hypothetical welfare-driven pro-recognition. The Tenth-Amendment objection to the preemption order raised by Sen. Joe Nicola (R-MO) is a marker that this resistance crosses partisan lines.
Argentina's automated-company proposal
Argentina's proposed "automated company" regime — announced by President Javier Milei in a June 2026 Financial Times op-ed as "non-human corporations" run by AI — would in practice require a human administrator to oversee operations and would keep companies liable for damage caused by their AI systems, according to the bill text detailed in July 3, 2026 reporting. The bill, co-authored by law professor Diego Duprat, also creates a DAO (decentralized autonomous organization) category requiring identified, registered token holders. Historian Yuval Noah Harari criticized the proposal over reduced corporate accountability (Source: reuters.com). The gap between the "non-human corporations" announcement framing and the bill's human-administrator requirement makes the proposal a legal-personhood datapoint of the agentic-AI-flexibility type: the regime extends corporate form to AI-operated entities while anchoring liability in human persons.
Relationships
- instance-of: AI Liability (the broader liability question, of which personhood is one structural piece)
- regulated-by: Missouri SB 1012 — Artificial Intelligence Non-Sentience and Responsibility Act (the first state-level attempt at a negative personhood definition)
- contradicts: AI Welfare / Model Welfare / Moral Patienthood (parallel but distinct thread on moral patiency — not strictly a contradiction but methodologically distinct)
- depends-on: Agentic AI (the agentic-AI-as-corporate-agent question is the operative industry-side concern)
- related: AI Federalism (the preemption dimension); Joe Nicola (the SB 1012 sponsor; stub)
Sources
- Politico Digital Future Daily, May 27, 2026 — Joe Nicola interview, "The case against AI personhood." (Source: politico.com)
- Missouri SB 1012 — Artificial Intelligence Non-Sentience and Responsibility Act — coverage of the SB 1012 bill, sponsor, kill mode, and Tenth-Amendment posture.
- AI Welfare / Model Welfare / Moral Patienthood — the parallel-but-distinct moral-patiency thread.