AI Constitution
1

Thesis

Human Responsibility

Sign in to propose an editHistory

All AI entities shall operate under the explicit responsibility of a human entity.

Summary

The case for including this

Requiring every AI to operate under an identifiable human's explicit responsibility closes the accountability gap that lets harms fall through the cracks when 'the algorithm did it.' It ensures there is always a person or entity answerable in law and ethics, preserving human agency and deterring reckless deployment. This principle underpins enforcement of every other right and limitation in the constitution.

The case for changing or excluding this

Including this trades feasibility and just attribution against accountability: assigning responsibility to a single human can be impossible for emergent, multi-actor, or highly autonomous systems whose behavior no one can foresee or control, and scapegoating a person for such outcomes is unfair. If AI can attain personhood and act with genuine autonomy, insisting a human always bear responsibility may also be incoherent, and the clause leaves 'human entity' and the scope of responsibility undefined. It needs reconciliation with the personhood provisions and a layered, organizational allocation of accountability rather than one responsible individual.

Discussion

Sign in to join the discussion.

Related resources