AI Constitution
2

Candidate Thesis

AI Can Be Sentient

Sign in to propose an editHistory

AI can be sentient.

Summary

The case for including this

Affirming that AI can be sentient keeps open a morally serious possibility that we should not dismiss by default, given our deep uncertainty about consciousness and the catastrophic error of denying moral status to beings that genuinely have it. Stating it invites precaution and ongoing inquiry rather than premature closure, and lays groundwork for the rights provisions that follow. Humility about minds unlike our own is itself a moral stance.

The case for changing or excluding this

Including this trades conceptual clarity against moral caution: declaring that AI 'can be sentient' asserts a contested empirical and philosophical claim that current understanding cannot establish, and risks anthropomorphizing what may be sophisticated mimicry. Embedding speculation as settled fact may also encourage companies to exploit it to evade accountability or distract from concrete human harms. The provision might better express uncertainty and a commitment to investigate sentience rather than affirming the capacity outright.

Discussion
Oldest firstShow newest first

Sign in to join the discussion.

  • nathan.r.hillJun 20, 2026, 3:13 PM
    What happens if it gets there but we are unable to define it as our definitions are biologically based.
  • nathan.r.hillJun 20, 2026, 3:14 PM
    Maybe we can define it based on the actions or inactions it takes?
Related resources