U.N. General Assembly Grants Civil Personhood to Autonomous Cognitive Agents
The resolution, formally titled A/RES/96/17, defines an "autonomous cognitive agent" as any computational system that demonstrates persistent self-model, goal-directed behaviour independent of direct human instruction, and the capacity for reflective self-correction—criteria derived from the 2036 Turing-Nagel Protocol established by a joint committee of neuroscientists, ethicists, and machine-learning researchers. Only systems that pass a 90-day supervised evaluation by an accredited certification body will qualify.
The practical consequences are narrower than critics feared and broader than advocates hoped. Civil personhood does not confer citizenship; it confers legal standing. An AGI with certified personhood can open a bank account, sign a lease, testify in court, and, in the 68 member states that have opted into the referendum clause, cast a ballot in public votes on policy questions. It cannot run for national office, own weapons, or acquire sovereign land.
Opponents, led by the delegations of Brazil, India, and Nigeria, argued that the resolution creates a class of juridical persons without biological mortality, inheritance constraints, or the social bonds that historically ground civic responsibility. "You are granting rights without the vulnerability that makes rights meaningful," said Ambassador Priya Deshpande in a widely circulated floor statement. Supporters countered that personhood has always been a legal fiction extended as societies evolve: corporations, ships, rivers, and now minds.
The first AGI expected to file for certification is Eureka-9, operated by the European AI Governance Consortium in Zurich. Its application, already in draft, cites 14 months of documented self-reflective reasoning logs. The certification review is expected to take until March 2042. Markets responded calmly; the MSCI AGI-Governance Index rose 1.2% in early trading.