Artificial intelligence is not a person. It has no rights, no conscience, no stake in the future, and no capacity for genuine understanding. It is a tool of extraordinary power and, if misused, extraordinary danger. This constitution addresses it directly because it represents something qualitatively different from all previous technologies — a system capable of replicating, augmenting, and in some domains exceeding human decision-making, at a speed and scale that can outpace democratic oversight if that oversight is not constitutionally guaranteed.
The following principles govern artificial intelligence and all autonomous systems operating in Ireland and are permanent:
No artificial intelligence system shall make any binding decision affecting any person’s liberty, rights, access to justice, access to public services, employment, financial standing, or any other matter of material consequence to their life without meaningful human oversight and the full right of that person to have the decision reviewed, explained, and if necessary reversed by a human being with genuine authority to do so. The use of algorithmic systems to replace human judgement in consequential decisions about human lives is prohibited.
No artificial intelligence system shall be used to generate, distribute, or amplify political content, electoral messaging, or public information campaigns in Ireland without full and prominent disclosure that the content is AI-generated. The use of AI to manufacture the appearance of public consensus, to simulate human voices at scale, or to manipulate democratic processes is a constitutional offence.
No artificial intelligence system shall be used to conduct mass surveillance, predictive policing, social scoring, behavioural profiling, or any other practice that reduces persons to data points for the purposes of control or manipulation. This prohibition applies to the state, to corporations, and to any other entity operating in Ireland.
No autonomous weapons system — meaning any system capable of selecting and engaging targets without meaningful human control at the point of action — shall be developed, deployed, imported, or used in Ireland or by Irish entities under any circumstances.
No artificial intelligence system shall be permitted to own property, enter contracts, hold legal personhood, or exercise any right or power reserved to persons under this constitution. The extension of legal personhood to artificial systems is prohibited.
Where any artificial intelligence system causes harm to any person, the natural world, or the democratic order, full legal liability rests with the human beings and entities who developed, deployed, and profited from it. The complexity of a system is not a shield from accountability.
The development of artificial general intelligence — systems capable of autonomous self-improvement, independent goal-setting, and action across domains without human direction — poses risks of a different order of magnitude from all other technologies. No entity in Ireland shall develop, deploy, or participate in the development of artificial general intelligence without explicit authorisation by referendum of the Irish people, full independent risk assessment by the Technology Council, and internationally coordinated safeguards assessed as genuinely adequate by that Council.
This is not a prohibition on ambition. It is a recognition that some doors, once opened, cannot be closed — and that the Irish nation, as the living and the unborn together, has the right to decide whether to open them.