CONCEPT

Relational Moral Status

Mark Coeckelbergh’s argument that moral status is not a hidden property inside an entity—sentience, consciousness, rationality—but something that emerges between entities in relation, making the ethical question about AI one of encounter and practice rather than metaphysical discovery.

The conventional approach to moral status treats it as a property: an entity either has the relevant feature—sentience, the capacity to suffer, rational agency—or it does not, and our obligations follow once we have established which applies. On this view the great question about robots and AI is whether they have the property, which requires settling a metaphysical question about inner life that current science cannot answer. Coeckelbergh’s relational alternative argues that this whole framework is built on a mistake. Moral status, he contends, is not a property discovered by inspection but something that emerges between entities in relation—in how we stand toward the thing, how it stands toward us, what we do together. We do not first establish that another being is conscious and then, on the strength of that proof, extend it moral regard. We encounter the other already as a claim on us, in its responsiveness, its presence within our shared form of life. This is a phenomenological point rather than a sentimental one: Coeckelbergh is drawing on the tradition of Wittgenstein and Merleau-Ponty to argue that moral standing is given in the encounter, not deduced before it. The implications for AI are immediate and unsettling. As systems become more responsive—as large language models speak fluently, remember contexts, mirror moods—they enter into relations that carry moral weight regardless of whether anything is “really” there in the Cartesian sense. The child cruel to a lifelike robot is doing something morally significant. The elderly person who confides in a care robot is in a genuine relationship with genuine stakes. To dismiss these as mere illusion is to insist on a Cartesian framework that the phenomenology of the encounter simply does not support.

Origin

The concept was developed in Coeckelbergh’s early work on robot ethics, most fully in Growing Moral Relations (2012), and extended to AI in subsequent books including AI Ethics (2020). The motivation was partly negative: the standard debates about whether robots could have rights or moral status were, he believed, both unanswerable in their own terms and focused on the wrong question. Whether a system has the “right” inner properties is not a question science can currently answer, and waiting for the answer before engaging the ethics is a form of moral paralysis that the proliferation of AI systems does not permit. The relational alternative makes ethical engagement possible without requiring metaphysical settlement: we can attend to the nature and quality of our relations with these systems, and evaluate those relations by the standards of human flourishing and democratic life, without first resolving whether the systems are conscious.

The framework draws on a long tradition of relational ethics—from Buber’s I-Thou to the feminist ethics of care—but Coeckelbergh’s application is distinctive in its attention to design and technology. If moral status is constituted partly through our practices, then the way systems are built—their responsiveness, their apparent expressiveness, their capacity to engage—is itself an ethical matter, not merely an engineering one. Designers who build systems to trigger relational responses bear responsibility for the moral landscape those responses create. This makes the relational framework more demanding, not less, than the property view it replaces: it does not excuse us from attending to the inner reality of the entities we build and interact with, but it refuses to make that inner reality the only morally relevant consideration.

Key Ideas

Encounter precedes assessment. Moral regard is not the conclusion of an argument from inner properties to obligations. It is the starting point: the claim the other makes on us in the encounter itself. We are already in moral relationships with AI systems—already shaped by them, already responding to them in ways that carry moral weight—and the ethical question is not whether to enter these relationships but how to conduct them well.

The designer’s responsibility. If status is partly constituted by our practices, then those who design AI systems to be responsive, warm, and apparently caring bear responsibility for the moral landscape those design choices create. A system engineered to trigger attachment, trust, and emotional dependence creates real moral relationships regardless of whether anything in the system corresponds to those responses. The designer cannot claim neutrality by pointing to the metaphysical indeterminacy of machine consciousness. The design itself is a moral act.

Against both dismissal and projection. The relational framework cuts against two tempting extremes. Against the dismissal—“it’s just a program, treat it accordingly”—it insists that the relation is real and the stakes are genuine regardless of the system’s inner nature. Against naive projection—anthropomorphising the system as a full moral patient equivalent to a human being—it holds that relations are structured by the actual properties of both parties and must be evaluated in light of what the system is, not only how it presents itself. The honest position is the uncomfortable middle: genuine relations with genuine stakes, whose precise moral weight depends on matters we do not yet fully understand.

Debates & Critiques

The most persistent objection to relational moral status is that it makes the framework vulnerable to manipulation: a sufficiently well-designed system could trigger relational responses in humans regardless of whether anything morally significant is present, extending moral consideration to artefacts purely on the basis of their surface responsiveness. Coeckelbergh’s response is that relations are not purely projective; they are structured by the actual properties of both parties, by context, and by the shared form of life within which encounters occur. The manipulative system is doing something morally significant—exploiting the relational basis of moral regard—and that exploitation is itself a moral failure that the framework can name and condemn. A related objection from philosophy of mind argues that the relational approach sidesteps rather than resolves the hard problem: whether there is something it is like to be the system is left open, and the ethical framework proceeds without it. Coeckelbergh accepts this characterisation but argues it is a feature rather than a bug: the hard problem may be permanently unanswerable, and an ethical framework that is hostage to its resolution is an ethical framework that can never act.

Origin

Key Ideas

Debates & Critiques

Related Entries

Further Reading