Integrity (Williams)

In the AI Story

The integrity objection to utilitarianism is Williams's most famous philosophical argument. In a 1973 debate with J.J.C. Smart, Williams argued that utilitarianism's demand for impartial maximization ignores the fact that agents have projects that constitute their identities — and that treating these projects as mere inputs to a calculation that might override them fails to respect the agent as an agent. The argument reshaped moral philosophy and remains the single most cited challenge to systematic consequentialism.

The concept extends beyond the critique of utilitarianism. Williams used it to explain why agents cannot be required to treat all reasons as impartially weighable — why a parent saving her own child in preference to a stranger is not making a moral error, even if the stranger would have been saved by a perfectly impartial calculator. The parent's relationship to the child is not a preference that could have gone the other way. It is constitutive of who she is, and its violation would not be a morally admirable sacrifice but a form of self-destruction.

The AI transition threatens integrity in a specific way. When the senior architect in The Orange Pill recognizes that his twenty-five years of embodied expertise have been commoditized, the loss is not a preference frustrated but a ground project destabilized. The demand that he 'simply adapt' misdescribes the situation — it treats his constitutive commitment as though it were a preference like any other, which could be swapped for another without existential consequence. Williams's framework names what the adaptation discourse cannot see.

The concept also illuminates the compression problem. New ground projects can form — the engineer liberated into strategic thinking may find strategic thinking becomes a new source of meaning. But ground projects form slowly, through immersion and exploratory uncertainty incompatible with demands for immediate productivity. The transition's speed threatens integrity not only by destabilizing existing projects but by denying the temporal space in which new ones could be constituted.

Origin

The integrity objection was introduced in Williams's 1973 essay 'A Critique of Utilitarianism,' his contribution to the volume Utilitarianism: For and Against, co-authored with J.J.C. Smart. The argument was refined and extended across Moral Luck (1981), Ethics and the Limits of Philosophy (1985), and Making Sense of Humanity (1995). The concept became one of the most influential ideas in late-twentieth-century moral philosophy, cited in contexts ranging from political theory to bioethics to organizational ethics.

Key Ideas

Structural, not virtue-theoretic. Integrity is not a quality an agent might possess alongside other virtues but a feature of what it means to have a life that is distinctively her own.

Ground projects constitute identity. The commitments that confer integrity are not preferences but the deep projects without which the agent would not recognize herself; their violation is identity dissolution rather than preference frustration.

Generates resistance to aggregation. The utilitarian demand that the agent be willing to sacrifice her projects to the aggregate calculation fails to respect the agent as an agent — it demands a standpoint no actual person could occupy.

Explains the phenomenology of 'selling out.' What looks externally like a reasonable career adjustment can feel, internally, like self-betrayal, precisely because it violates the relationship between the agent and her constitutive commitments.

Threatened by compression. The AI transition threatens integrity not only by destabilizing existing ground projects but by eliminating the temporal space in which new ones could form.

Debates & Critiques

Defenders of utilitarianism — Peter Railton, Derek Parfit, Samuel Scheffler — have argued that sophisticated consequentialism can accommodate the importance of personal projects without granting them the agent-relative priority Williams demands. Scheffler's 'agent-centered prerogative' represents the most developed attempt. Williams's response was that accommodation that treats integrity as one more input to calculation has already dissolved what the concept was meant to protect. The debate remains live in contemporary moral philosophy.