If postmodern theory undermines claims to objectivity, universality, and certainty, what does it leave in their place? Does it leave anything, or does it open into a void of relativism in which no ethical claims can be made?

This is the challenge that postmodern theory must answer if it is to contribute constructively to AI ethics. And postmodern thinkers, as well as scholars who engage critically with postmodernism, have developed responses.

Situated ethics. Rather than applying universal principles to abstract individuals, situated ethics attends to the specific context, relationships, and consequences of particular decisions. For AI, this means asking not “is this AI system fair?” in the abstract but “for whom is it fair, in what context, with what consequences, and who bears the cost of its errors?” This is more demanding than abstract principle-application. It requires genuine engagement with the specific people affected by specific systems.

Epistemic humility. Postmodern insights about the limits of knowledge do not imply paralysis. They imply humility, designing AI systems with awareness of what they do not know, building in mechanisms for correction when they are wrong, resisting the authority of algorithmic outputs when that authority is not warranted. Epistemic humility is, in fact, a form of rigour, a more honest accounting of uncertainty than false confidence in model accuracy.

Plural and provisional. Rather than designing AI systems around a single model of the good, postmodern-influenced AI ethics embraces plurality, the coexistence of multiple values, frameworks, and ways of knowing. Rather than presenting outputs as final determinations, it treats them as provisional, subject to revision, contestation, and override. This is not a recipe for indecision. It is a recipe for accountability.

Contestability as a design principle. If AI systems construct realities rather than merely describing them, then the people subject to those constructions must have genuine power to contest them, not merely to request explanations, but to challenge the standards, assumptions, and data on which AI decisions are based. Contestability, understood in this fuller sense, is a postmodern contribution to AI governance design.

The ethical is political. Postmodern analysis ultimately insists that AI ethics cannot be separated from AI politics, from questions of power, ownership, accountability, and democratic governance. The attempt to resolve ethical questions through technical means, while leaving political questions untouched, is itself a political act, one that serves the interests of those who benefit from the current organisation of AI development and deployment.

Living with uncertainty is not comfortable. But it is honest. And in a field that has caused significant harm through overconfident claims about what AI systems know and can do, honesty is among the most important ethical commitments we can make.