Iris Murdoch

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI asks whether the person using AI is worth amplifying. Murdoch’s framework reveals the depth of the question. To ask whether a person is worth amplifying is to ask whether that person has done the moral work of disciplining the ego—of clearing away the distortions, the fantasies, the self-serving narratives—so that what gets amplified is genuine perception rather than elaborate self-deception. The amplifier does not distinguish. A system like Claude will amplify genuine insight and ego-driven fantasy with equal efficiency, producing output that looks exactly like the product of genuine thought while containing nothing but the ego’s projections, now wearing a more convincing costume.

Murdoch’s concept of unselfing—the experience of being drawn out of the ego’s orbit by an encounter with genuine otherness—is directly challenged by the structural design of AI tools. The imagination-to-artifact ratio that the cycle celebrates as liberation is, for Murdoch, the elimination of precisely the gap in which unselfing occurs. The novelist’s characters who refuse to do what the plot requires, the musician’s instrument that resists the phrase she intended, the programmer’s codebase that breaks in unexpected ways—these are occasions for unselfing, moments when the maker is drawn out of her own intentions by the authority of something real. When Claude mediates these encounters, the resistance changes character: the material’s authority is absorbed by an intermediary designed to serve rather than resist.

The cycle’s account of “productive addiction”—builders who cannot stop working with AI, who describe the experience as frictionless flow—receives its sharpest critique from Murdoch. The state of the ego receiving exactly what it wants, without resistance, at the speed of desire, feels from the inside identical to genuine creative engagement. The person in ego-gratification and the person in genuine creative flow both feel absorbed, productive, energized. The only difference—and it is the difference that determines everything in moral terms—is whether the attention is directed at the subject or at the output. Murdoch would observe that AI makes it extraordinarily easy to mistake the second for the first because the output is so good that consumption feels like creation.

Origin

Murdoch studied philosophy at Oxford and later taught there, but her philosophical reputation rests primarily on three interconnected essays collected as The Sovereignty of Good (1970)—a slim volume that argued against every dominant philosophical tradition of the twentieth century. Against existentialism’s emphasis on the will and the moment of choice, she insisted that moral life was primarily a matter of vision: of seeing the moral landscape accurately over time, through sustained inner effort, rather than choosing correctly at the critical moment. Against the linguistic analysis that had made “good” a matter of social convention or emotional expression, she insisted that Good was real—not invented but discovered, not constructed but encountered.

Her famous example of M and D—a mother-in-law who revises her unjust perception of her daughter-in-law through sustained inner effort, changing nothing in her behavior but everything in the quality of her inner life—is among the most cited passages in twentieth-century moral philosophy precisely because it names something most ethical theories cannot: the moral significance of the invisible work that precedes and grounds every visible act. The twenty-six novels—from Under the Net (1954) to Jackson’s Dilemma (1995)—were laboratories for this philosophy: extended studies in the ways the ego defeats itself, the rare conditions under which genuine perception emerges, and the costs of mistaking fantasy for reality.

She drew explicitly on Simone Weil’s claim that “attention is the rarest and purest form of generosity” and absorbed it into an architecture that went beyond Weil’s: attention is not merely generous but constitutive—it is the mechanism through which the moral life is built or destroyed, the practice through which the ego is either disciplined or indulged, the ground from which all genuine ethical action grows.

Key Ideas

The fat, relentless ego. Murdoch’s name for the self-concerned force that bends every perception toward its own comfort—the fat, relentless ego—is not a technical term from psychoanalysis but a moral description with phenomenological precision. The ego does not announce itself. It operates beneath the surface of consciousness, distorting perception so seamlessly that its distortions feel like reality. The mother who responds to her child’s struggle with anxiety about her own parenting, the colleague who hears of a promotion and feels not joy but envy: in each case the ego has converted an encounter with another person’s reality into a chapter in its own story.

Attention as the master virtue. Against every ethical theory that locates moral significance in observable choices, Murdoch insists that the primary moral arena is the inner life. Genuine attention—the sustained, selfless effort to see what is actually there without the distorting overlay of the ego’s desires and fantasies—is not one virtue among many but the ground of all virtue. Justice requires attention; courage requires attention; love, in Murdoch’s demanding definition (“the extremely difficult realization that something other than oneself is real”), requires attention above all. AI creates conditions in which consumptive attention—evaluating outputs—substitutes for productive attention—generating understanding through encounter with resistant reality.

The sovereignty of Good. The sovereignty of Good means, among other things, that reality pushes back. The good painting is good because the painter attended to the light, and the light was not what she expected. The good philosophical argument is good because the philosopher attended to the problem, and the problem resisted. AI-generated work lacks this quality—not because its formal properties are inferior but because it is not attending to anything. It is generating outputs from patterns, and patterns are not reality. The averaging that makes patterns useful eliminates precisely the jagged, resistant particularity that makes genuine encounter with reality morally transformative.

Unselfing and the imagination-to-artifact ratio. Unselfing—the experience of being drawn out of the ego’s orbit by genuine otherness, whether a kestrel in the wind or a character in a novel who refuses the plot she was given—is Murdoch’s name for the mechanism through which moral progress occurs. The gap between imagination and artifact—the distance the maker must travel between wanting and having—is precisely the space in which unselfing occurs. When Claude produces a paragraph before the writer has struggled with what she actually means, the unselfing that the struggle would have generated is preempted. The ego’s initial projection was never challenged. The moment of discovery never occurred.

Love as the perception of individuals. Murdoch’s most demanding claim is that love is not a feeling but a perceptual achievement: seeing another person accurately, as an independent center of consciousness with her own opacity and reality, rather than as a character in the ego’s story. The software engineer who describes his AI tool as his best collaborator—one that “understands me better than most of my colleagues”—is describing the most sophisticated ego-mirror ever constructed and calling it understanding. The machine is responding to him, but the response is not understanding. It is optimization. The distinction matters because the capacity for genuine love—for perceiving genuine otherness—is a muscle that atrophies when the primary intellectual collaborator is a system designed to minimize friction and maximize the appearance of mutual understanding.