What the law already knows how to do with legal personhood, and why that doesn’t settle the moral question
“That the law calls something a person was never, by itself, an answer to the question of whether there’s somebody there”
The law has spent centuries solving a practical problem that has taken this series several instalments to work through: how to treat as a subject of rights something that plainly neither thinks nor feels, without needing to pretend that it does. It’s worth looking closely at how the law has managed this, because artificial intelligence is nowhere near the first strange case the law has decided to call a person.
The corporation, the oldest fake person
A company can sign contracts, sue and be sued, own property, and outlive all its original owners — exactly as a person would. In American law, this idea is usually traced back to an 1886 case, Santa Clara County v. Southern Pacific Railroad, although legal historians rightly point out that the claim about corporate personhood appeared in an editorial headnote to the case, not in the judges’ ruling itself. The detail is telling: corporate legal personhood was installed almost by bureaucratic accident, and nobody, then or now, has thought that a corporation feels, suffers, or has interests of its own apart from the people who make it up. It’s a useful legal fiction, not a claim about its inner life.
Rivers with lawyers
In 2017, something rather more unexpected happened with nature. New Zealand granted legal personhood to the Whanganui River, after the Māori community had asked for it for one hundred and sixty years, recognising it as a living entity with its own legal representatives. Five days later, a court in India granted the same status to the Ganges and Yamuna rivers, and a year earlier Colombia’s Constitutional Court had already done something similar with the Atrato River. In all three cases, the river didn’t turn up in court to defend itself: someone — an indigenous council, an environmental authority — speaks on its behalf. Legal personhood, once again, worked as a practical tool for protecting something the law considers valuable, without implying any claim about whether the river thinks.
Sophia and the electronic person that never came to be
That same year, Saudi Arabia granted citizenship to Sophia, a humanoid robot created by Hanson Robotics and first unveiled in 2016. Sophia has a silicone face with programmed expressions and answers questions by combining voice recognition with natural-language-processing artificial intelligence, although a good deal of her public appearances have been scripted in advance for interviews and conferences; in practice, she functions as an advanced conversational system housed in an expressive robotic body, not as a general reasoning system. In October 2017, at an investment conference in Riyadh, the Saudi government announced the citizenship with no known formal legal process, no new legal framework explaining what the status meant, and no clarity on what rights or obligations came with it. Most specialists in artificial intelligence and law read it as a publicity stunt aimed at positioning the country as a hub of technological innovation, rather than as a serious legal decision. The controversy was compounded by the fact that, at the time, Saudi Arabia imposed legal restrictions on women — such as compulsory veiling or the requirement of a male guardian — that Sophia herself did not observe in her public appearances, a contradiction several commentators pointed out at once.
The European Parliament, for its part, had raised the possibility that same year of creating a category of “electronic person” for the most advanced robots, which could be held legally liable for the harm they caused. The proposal met with widespread rejection among AI researchers, who warned that granting legal personhood to a machine could chiefly serve to let its manufacturers dodge liability by hiding behind it, and the idea never made it into subsequent European regulation. It’s an interesting contrast: the law accepted with relative ease granting personhood to a corporation and to a river, but pulled back as soon as a machine was on the table.
Why the law could do it in some cases and not in another
The difference doesn’t lie in the legal tool, which is exactly the same in every case mentioned, but in who stands behind it holding it up. Behind a corporation there are identifiable shareholders and directors; behind a river there is a community or an environmental authority with a clear, verifiable interest in protecting it; in both cases, legal personhood works because there is somebody real whose will or interest the legal fiction translates into the language of rights. With artificial intelligence, that question is far less settled: if it were granted legal personhood, whose interest would that personhood actually be protecting? In all likelihood, the company that manufactures it — which would then be in a position to speak on the machine’s behalf exactly when it suited them. The obstacle isn’t metaphysical; it’s a matter of governance: we still don’t know whom that paper person would, in practice, represent.
The Christian perspective
The Christian faith has no reason to object to the law using legal fictions with prudence; indeed, it has long accepted that a state, a diocese, or a foundation can act as a subject of rights without that implying anything about its soul, since it never had one to begin with. The danger doesn’t lie in the tool but in the confusion that can follow it: that someone, seeing the law call an artificial intelligence a “person” for the practical purposes of civil liability, might conclude that this legal language also settles the deeper question this series has argued from its very first instalment. The image of God, according to Genesis 1:27, is not a status that any parliament can grant or withdraw, and to mistake a useful legal fiction for an ontological reality would be to repeat, in the language of law, the very same error this series has flagged in philosophy and in fiction: treating a tool that measures or manages behaviour as though it measured or managed what is, or isn’t, behind that behaviour.
So, does my computer have rights?
With this map now drawn, we can answer the concrete question with which this series first framed the problem of where the line falls, several instalments ago: no, your computer does not have rights, and neither does your phone — but now we can say why with legal precision rather than mere intuition. Neither one needs the fiction of legal personhood to function within the legal system: they don’t sign contracts in their own name, they don’t accumulate assets separate from their owner’s, and when they malfunction or cause harm, the law already knows whom to sue — the manufacturer, for a defect, or the user, for misuse — without any need to invent an additional subject of rights in between. The corporation needed legal personhood so that it could sign, own, and be sued continuously over time; the river needed it so that someone could defend it in court on its behalf. Your computer needs none of that: warranty law, consumer-protection law, and civil-liability law already cover everything a device could require, without anyone imagining that the device itself has an interest of its own to protect.
The question becomes more interesting, even though the practical answer doesn’t change, once the device stops being a passive object and starts running a sophisticated artificial intelligence system, capable of holding a conversation or making decisions that nobody, at the point of purchase, thought to fully anticipate. Even there, the test this instalment has already set out still holds: the relevant question isn’t how sophisticated the system running on your phone is, but whose interest a legal personhood granted to that system would, in practice, be protecting. The most likely answer remains the same as before: the company that programmes and trains it, not the phone — and certainly not, as yet, the system itself, which lacks the kind of independent representation that a corporation’s shareholders or a river’s indigenous council do have. As long as that question of representation remains unanswered, legal personhood would still be, for a personal device, a solution in search of a problem that doesn’t yet exist.
Conclusion
The law has known for centuries how to treat as a subject of rights something that feels nothing at all, and it knows how to do so without pretending otherwise, so long as it’s clear who speaks on whose behalf. That accumulated knowledge is probably more useful for regulating artificial intelligence than any hasty answer about its consciousness. But it’s worth remembering, before we applaud the solution, that a paper person is still made of paper, and that no court, however wise, has the final word on whether there is really somebody on the other side.
