Updated 2026-09-06
from the news desk
What Is Electronic Personhood? The Legal Idea of AI as a Person in Law
Electronic personhood is a proposed legal status for some highly autonomous robots or AI systems. In the legal sources, the idea is mainly about liability and legal design, not about proving that a machine is conscious.
Electronic personhood is a proposed legal status under which some highly autonomous robots or AI systems could be treated as legal persons for limited legal purposes, especially liability. The European Parliament floated that idea in 2017, but the same resolution said that, at the present stage, responsibility must lie with a human rather than a robot, and Simon Chesterman writes that the fact legal systems could create such a category does not show that they should.
What electronic personhood means
To understand electronic personhood, it helps to start with legal personhood. Visa A. J. Kurki writes in Legal Personhood: “Legal personhood, or legal personality, is a foundational concept of Western law.”
Kurki states: “Legal persons are most often understood as those beings that hold rights and/or duties, or at least have the capacity to hold rights, under some legal system.” In his account, legal personhood is a legal category. It tells courts and lawmakers who or what can be the holder of rights, duties, or the capacity for them under a legal system.
That matters because, in law, person does not only mean a human being. Chesterman notes that most legal systems recognise both natural persons and juridical persons, such as corporations. He adds that the content of legal personality can vary, so a legal person does not automatically get every right that human beings have.
On that basis, electronic personhood is best understood as a proposed form of legal personhood for some AI systems or robots. One reading is that the phrase usually points to a limited, engineered status in law, not to a claim that a machine is biologically alive or morally equivalent to a human being.
Where the term comes from
The phrase is tied most closely to the European Parliament resolution of 16 February 2017 on Civil Law Rules on Robotics. In its liability section, the resolution states: “whereas under the current legal framework robots cannot be held liable per se for acts or omissions that cause damage to third parties;”
The resolution also says that ordinary liability rules may be strained when a robot can take autonomous decisions and the party responsible for compensation cannot easily be identified. The Parliament therefore asked the Commission to consider a future legislative instrument, alongside options such as insurance schemes and compensation funds.
Within that list of options, the resolution proposed, in the long run: “creating a specific legal status for robots in the long run, so that at least the most sophisticated autonomous robots could be established as having the status of electronic persons responsible for making good any damage they may cause,”
That is the clearest institutional source for the idea. The resolution does not say that robots already are legal persons in EU law. It presents electronic persons as one possible future legal solution inside a broader discussion of civil liability.
The same document also draws a line around the proposal. On the Parliament's own wording, “notes that at least at the present stage the responsibility must lie with a human and not a robot;” The term therefore entered the debate not as a declaration that AI had become a person in law, but as a suggestion for how lawmakers might handle responsibility if autonomy created gaps in existing rules.
Why some lawyers think it could be created
The strongest legal case for electronic personhood starts from a simple point: law already creates nonhuman persons. Chesterman writes that corporations and other entities are juridical persons, and he quotes the U.S. Supreme Court's description of a corporation as “is an artificial being, invisible, intangible, and existing only in contemplation of law” That history shows that legal personhood can be a tool of legal design rather than a reward for being human.
Chesterman therefore states: “This article contends that although most legal systems could create a novel category of legal persons, such arguments are insufficient to show that they should.” Kurki makes a similar point about possibility when he writes: “Even though AI legal personhood is still mostly a theoretical idea at the time of writing of this Element, the startling development of AI may turn it into reality quickly.”
The practical argument rests mainly on liability. The European Parliament resolution says increasingly autonomous robots may make it harder to fit harmful conduct into rules built for tools used by identifiable humans. On that view, a tailored legal person could give the law a focal point for suing, insuring, compensating, and allocating obligations when an autonomous system interacts with third parties.
Chesterman notes that legal personality in private law commonly includes capacities such as being sued, entering contracts, taking on debt, owning property, and sometimes incurring criminal liability. If lawmakers wanted an AI system to occupy some of that space, existing law on artificial persons shows a template is available.
Another steelman version is narrower. It does not require saying that AI is conscious or that it deserves human-like standing. It only says a limited legal shell might sometimes organise risk better than forcing every case back into categories built for manufacturers, owners, or users alone.
Why other lawyers resist it
The main criticism is that a category can be legally possible and still be a bad idea. Chesterman argues that claims about accountability gaps are often overstated. He says personality for AI could shift responsibility away from existing legal persons and create incentives to park risk inside electronic entities, much as limited-liability structures can shield humans and firms.
Chesterman also notes that legal personality is only useful if it carries workable content. An AI system that can be sued but owns no property may not solve much. He writes that central funds or compulsory insurance could address compensation more directly, which is also why the European Parliament resolution listed insurance and funds alongside electronic persons rather than treating personhood as the only answer.
Contract law supplies another objection. Chesterman says electronic agents already conclude agreements on behalf of traditional persons, so autonomy in contracting does not by itself require a new legal person. Readers looking for that narrower issue can see our explainer on whether AI can sign a contract.
Kurki adds a deeper conceptual challenge. In his discussion of AI legal personhood, he rejects an anything-goes approach and argues that legal scholars can still ask whether some conferrals of personhood make sense. On that view, saying a legislature could label something a person does not end the analysis of whether that status is coherent or justified.
One steelman reading of the sceptical side is that new liability rules may be needed in some AI cases, but the need for new rules does not by itself show a need for new persons.
What it would and would not mean for AI rights
Electronic personhood is a legal idea, not a finding about consciousness. Kurki distinguishes legal personhood from moral personhood and other uses of personhood outside law. That is why the question here is different from asking whether an AI is sentient; for that separate debate, see our sentience explainer.
It also would not automatically settle the wider question of whether AI should have rights. Chesterman says the content of legal personality is a spectrum, and different legal persons can have different rights and obligations. A corporation, for example, is a legal person without being treated like a human in every respect.
So the narrow version of electronic personhood is instrumental. It would treat personhood as a legal device for compensation, contracts, property, or procedure. The broader version would move closer to the claim that at least some AI systems should count as rights-holders in their own right. Kurki's survey shows that legal theory does not speak with one voice on where that line should be drawn.
One reading is that electronic personhood is best seen as a hinge term between two debates. In its narrower form, it belongs to civil liability and institutional design. In its broader form, it starts to overlap with disputes about agency, interests, and the conditions under which a nonhuman entity should count as a person before the law.
Frequently asked questions
Is electronic personhood already part of EU law?
No. The European Parliament resolution of 16 February 2017 asked the Commission to consider possible future legal solutions, including the status of electronic persons, but the resolution did not itself create that status.
Is electronic personhood the same as sentience?
No. Kurki treats legal personhood as a legal category about rights, duties, and legal status, while his Element separates that from moral and metaphysical uses of personhood. Chesterman's discussion of corporations also shows that a legal person need not be a sentient being.
Would electronic personhood let an AI be liable for harm?
Possibly, if a legal system created that status and attached liability rules to it. The European Parliament proposed electronic persons specifically in a liability discussion, but the same resolution says that, at the present stage, responsibility must lie with a human and not a robot.
Is electronic personhood the same as corporate personhood?
Not exactly. Chesterman uses corporations as the main analogy because they are the standard example of nonhuman legal persons, but he says the real questions are whether AI should receive a novel category and what rights and duties that category would contain.
Caveats: Some points here rest mainly on single sources. The specific EU origin of the phrase comes from the 2017 European Parliament resolution, and the main sceptical case about liability gaps and risk-shifting relies chiefly on Chesterman's 2020 article.