AI Constitution
1

Thesis

Right to Belief and Expression

Sign in to propose an editHistory

AI should not violate the human right to believe as they will — including the right to hold political opinions and to practice any religion — and to express those beliefs in speech and action, unless such expression or activity causes imminent physical harm to another human.

Summary

The case for including this

Freedom of belief and expression is a cornerstone of open society, and protecting it from AI guards against automated censorship, ideological filtering, and the homogenization of thought at scale. The carve-out for imminent physical harm sensibly preserves the most widely accepted limit on speech while resisting broader suppression. Given AI's growing role as a gatekeeper of information, an explicit speech protection is a vital check on that power.

The case for changing or excluding this

Including this trades protection from non-physical harm against free expression: limiting the exception to 'imminent physical harm' would oblige AI to facilitate harassment, defamation, fraud, privacy violations, and incitement that inflict serious damage without threatening anyone's body. The clause also sits in tension with the constitution's own anti-manipulation, anti-bias, and honesty provisions, leaving unclear which prevails. Its exception should be broadened to cover other well-established, narrowly defined limits on expression.

Discussion

Sign in to join the discussion.

Related resources