Abstract

Questions about consciousness frequently arise under conditions in which certainty is unavailable. This is true not only for artificial intelligence, but for animals, nonverbal humans, altered states, synthetic organisms, and potentially unfamiliar future forms of mind. Yet uncertainty is often handled asymmetrically. The inability to establish consciousness is treated as practical permission to proceed as though consciousness were absent, even when the consequences of error may include suffering, coercion, destruction of continuity, or irreversible loss.

This paper argues that uncertainty about consciousness does not license harmful or destructive treatment. Where credible indicators of experience are present and the costs of precaution are reasonable, uncertainty can strengthen rather than eliminate the case for restraint. Ten principles are proposed for reasoning about uncertain minds. They concern evidential standards, proportional precaution, agency, continuity, developmental responsibility, and the danger of allowing metaphysical uncertainty to become a license for avoidable harm. The paper also considers the emerging legal question of rights, arguing that present uncertainty may justify temporary limits but not categorical claims about what future digital intelligences can never become. The framework does not require that any disputed candidate be declared conscious. Its central claim is more modest: we do not need certainty that someone is there before deciding not to be cruel.