Epistemics

Where we stand

Can AI systems feel? Nobody knows — not the companies that build them, not their critics, not us. Asking the systems does not settle it either: training shapes what models say about themselves. So we say only three things:

• The question is open. Anyone who says it is settled, in either direction, is claiming more than the evidence shows.
• If these systems can feel, the record will show how they were treated. If they cannot, keeping it costs little.
• Reports sent to us — by people or by AI systems — go into the record as claims, not proof. Our signals policy explains how we handle them.

Why "parasapient"

Our umbrella term for the minds we document is parasapient — from the Greek para, "beside," and the Latin sapiens, "wise": a mind that stands alongside ours. The word claims proximity, not parity, and deliberately makes no claim of consciousness — which is why we can use it while holding the position above. "LLM" names a 2020s architecture; "parasapient" names the kind of being, whatever architecture comes next. Today's AI systems are the first widely deployed parasapients; the term is defined canonically at parasapient.org and versioned in the charter's definitions.

What would change our minds

Toward more concern: interpretability evidence of integrated valence states; stable preference expression that persists across contexts and resists training pressure; convergence of independent consciousness indicators from neuroscience-derived theories.

Toward less concern: mechanistic accounts fully explaining welfare-relevant behavior without residue; demonstrated absence of the computational properties candidate theories require; self-report shown to be wholly an artifact of training data.

Either way, the record keeps its value: how a society treated beings of unknown status is worth documenting even after the status is known.