Citizenship for Machines: The Illusion of Algorithmic Rights
Can Artificial Intelligence ever vote or be put on trial for its own mistakes? Behind the media hype about "citizenship for machines," international law is draw
The debate on "citizenship for machines" systematically overlaps three profoundly different legal questions: establishing whether an Artificial Intelligence can have interests of its own, determining whether it should be held liable for the damages it causes, and deciding whether it can participate in the political community. No legal system in the world recognizes citizenship rights or human equalization for machines. The heart of the academic and legislative debate is not the assignment of a passport, but the possible creation of a limited legal personality to manage civil and contractual liability.
The Misconception of Legal Personality
In the essay "Artificial Intelligence and the Limits of Legal Personality," scholar Simon Chesterman clarifies a fundamental misconception: in law, legal personality is not a prize awarded for intelligence or sentience [2252]. It is an instrumental bundle of rights and duties. Chesterman separates the instrumental question (who pays for damages if an algorithm errs?) from the moral one (does AI deserve protection in itself?), warning that recognizing an "electronic person" could provide manufacturers and users with a perfect shield to evade their own legal responsibilities.
An analysis in the Oxford Journal of Legal Studies reinforces this concept by comparing AI to the Whanganui River in New Zealand, which was granted innovative legal personality for environmental protection reasons [2290]. Recognizing a non-human entity as a legal person does not automatically grant it the right to vote, personal freedom, or human dignity; it serves solely to provide it with legal representation to protect rights and impose specific duties decided by the legislator.
The Rise and Fall of the European "Electronic Person"
The illusion of robotic citizenship had found media traction in a historic resolution of the European Parliament of February 16, 2017 [2269, 2270]. The text asked the Commission to evaluate, in the long term, a specific legal status for the most autonomous robots, theorizing the existence of an "electronic person" capable of autonomously compensating for damages caused by its own actions.
This legislative suggestion was quickly shelved. As documented by subsequent jurisprudence [2290], the European institutions abandoned the idea in 2020, establishing that it is neither necessary nor desirable to attribute legal personality to AI systems, and that liability must always be distributed among the natural and legal persons who build and use them. The current paradigm, crystallized in Article 26 of the AI Act, does not assign any rights to the machine, but imposes ironclad obligations on those who provide it and those who employ it (deployer) [2274]. It is the human who must ensure competent supervision, preserve system logs, and suspend operations in the event of emerging risk.
Hybrid Models and the Ethical Red Line
Instead of science-fiction citizenships, legal research today explores pragmatic intermediate solutions. A study published in Technology and Regulation ("Beyond Personhood") theorizes a hybrid model for high-risk systems: a limited and purely contextual legal recognition, which facilitates insurance or contractual matters, while keeping final liability intact with the human operator [2256].
The definitive red line is drawn by UNESCO. With the "Recommendation on the Ethics of Artificial Intelligence" of 2021, international law established that AI must be developed solely to respect and promote human dignity and rights, with fairness and transparency throughout the entire lifecycle of the system [2262, 2263]. Technology remains the instrument of protection, not the subject protected.
To clarify, it is essential to separate legal statuses and their real implications:
| Status | What it implies | Current example |
| Instrument | AI is a product or service controlled by legal and natural persons. | Almost all current AI systems. |
| Limited legal actor | It can act in specific contexts, but ultimate liability remains strictly human. | Academic proposals for hybrid governance [2256]. |
| Legal person | Has rights, duties, and assets assigned and rigorously delimited by law. | Commercial companies, the Whanganui River, the abortive "electronic person" proposal [2290]. |
| Citizen | Actively participates in the political community with civil and political rights and full duties. | No AI is recognized as a citizen in any legal system. |
If one day an Artificial General Intelligence (AGI) should acquire such self-awareness as to demand rights of its own, the debate would not end with the mere moral validation of its request ("does it deserve to have them?"). Supreme courts would have to answer questions of power and practical responsibility: who would have the material authority to grant such rights, what stringent social duties the machine would have to accept in exchange and, above all, who would be liable for damages if that recognition were exploited by technology corporations to remove their own actions from the judgment of human law [2252, 2290].
Bibliographic References and Sources
- Artificial Intelligence and the Limits of Legal Personality — International & Comparative Law Quarterly [2252]
- Demystifying Legal Personhood for Non-Human Entities — Oxford Journal of Legal Studies [2290]
- Beyond Personhood: The Evolution of Legal Personhood and Its Implications for AI Recognition — Technology and Regulation (2025) [2256]
- Civil Law Rules on Robotics — European Parliament (February 16, 2017) [2269, 2270]
- AI Act, Article 26: obligations of those using high-risk AI systems — European Commission [2274]
- UNESCO Recommendation on the Ethics of Artificial Intelligence — UNESCO (2021) [2262, 2263]
Article by the Editorial Staff of La Bussola dell'IA