Key takeaways
- Legal personhood is a tool, and has been granted to non-human entities such as corporations and a river in New Zealand.
- Courts and regulators currently treat AI as a product, with responsibility resting on people and firms.
- Whether machines can be conscious is unresolved and separate from legal questions about liability.
In 2017, New Zealand’s parliament recognised the Whanganui River as a legal person, able to be represented in court. In the same year, Saudi Arabia granted “citizenship” to a robot called Sophia, a publicity stunt that disturbed many lawyers. Between those two events lies a real question: when should the law treat something that is not a human being as a bearer of rights or duties?
Personhood is a tool
Legal personhood has never meant being human. Corporations, the first widespread example, can own property, sign contracts and be sued, because it is convenient for society. Ships have been “defendants” in maritime law for centuries. The question is not whether a machine is like us, but whether treating it as a legal subject solves a practical problem.
The problem of responsibility
The practical problem is accountability. In March 2018, an autonomous test vehicle struck and killed a pedestrian, Elaine Herzberg, in Arizona. The case raised the question of who was at fault: the company, the engineers, the backup driver, or the system. Prosecutors eventually charged the safety driver. Most current legal systems, including the European Union’s AI Act, place responsibility on developers, deployers and users, not on the system itself.
The question of mind
Rights are often justified by the ability to suffer or to have interests. Alan Turing proposed in 1950 that whether a machine “thinks” should be judged by its behaviour. John Searle replied in 1980 that a machine following rules need not understand anything, an argument that remains debated. No existing system is accepted by scientists as conscious, though there is no agreed test.
Editorial assessment
For now, the strongest argument is against giving machines legal responsibility: it risks giving the people who build and profit from them a way to point at the machine. Responsibility should follow control and benefit. Whether rights for artificial minds will ever be warranted is a question for a future we cannot yet describe. Most of the urgent questions concern the humans who design and deploy these systems today.
Questions to think about
- If an autonomous system causes a death, who should answer for it?
- Would granting rights to machines help victims or let developers escape accountability?
Disagree? Say why in the discussion below. Good arguments change minds.
References and sources
- Alan Turing, “Computing Machinery and Intelligence,” Mind (1950)
- John R. Searle, “Minds, Brains, and Programs,” Behavioral and Brain Sciences (1980)
Discussion
Challenge the argument, not the person. Community guidelines