---
title: "The Five Stages Of AI Grief"
description: "In this June 2024 Noema essay Benjamin Bratton proposes that artificial intelligence is a Copernican trauma in Freud's sense, a decentering of the human from a position of intuitive privilege, and…"
type: "reading-notes"
authors:
  - "Benjamin Bratton"
published: "2024-06-20"
original_url: "https://www.noemamag.com/the-five-stages-of-ai-grief/"
canonical_url: "https://antikythera.wiki/work/media/noema-the-five-stages-of-ai-grief"
md_url: "https://antikythera.wiki/md/work/media/noema-the-five-stages-of-ai-grief"
last_updated: "2026-08-23"
site: "Antikythera Wiki"
---

# The Five Stages Of AI Grief

> Independent, unofficial reading notes from Antikythera Wiki (https://antikythera.wiki). Written with AI from the published text, with citations; not by the work's author and not affiliated with Antikythera (https://antikythera.org). Read the original at the link below.

- **Authors:** Benjamin Bratton
- **Published:** 2024-06-20
- **Kind:** Essays, talks & interviews · Essay
- **Venue:** Noema Magazine
- **Original:** https://www.noemamag.com/the-five-stages-of-ai-grief/
- **Length:** 5.5k words
- **This page:** https://antikythera.wiki/work/media/noema-the-five-stages-of-ai-grief
- **Notes generated:** 2026-08-23

## Summary

In this June 2024 Noema essay [Benjamin Bratton](https://antikythera.wiki/people/benjamin-bratton) proposes that artificial intelligence is a Copernican trauma in Freud's sense, a decentering of the human from a position of intuitive privilege, and that Western responses to it can be sorted with Kübler-Ross's five stages of grief: denial, anger, bargaining, depression, and acceptance. Each stage is treated as a genre of argument with its own symptomatic statements, its own kernel of insight, and its own self-deception. The essay then steps outside the typology to describe a "non-grief" position, drawn from Sara Walker's account of life and technology as one evolutionary process and from James Lovelock's Novacene, in which machine intelligence is a phase in the long complexification of intelligence rather than a threat to it. It is Bratton's most compact account of why he thinks the current AI debate is inadequate to its subject, and a bridge between the planetary-computation program and the later Agentworld work.

## The argument

Bratton opens with Ilya Sutskever burning a wooden effigy of "unaligned" AI at an OpenAI retreat, read as a reaction to a technology that disturbs the human self-image. The conceptual frame comes from Freud, who called Copernicus, Darwin, and psychoanalysis "traumas" of decentering. Bratton adds neuroscience, biotechnology, and now AI. The trauma is not the technology but what it reveals: that intelligence, cognition, and mind are not what they feel like and not uniquely human. Human sapience has artificialized shelter, food, energy, and images; now intelligence is itself artificializable.

Trauma brings grief, and grief has stages. Bratton is explicit that the typology is not a diagnosis and that the stages do not run in order; denial and anger routinely coexist in critics who say AI is not real and must be stopped. Each stage is adopted because it flatters the griever or fits prior commitments, and all of them, he argues, miss the most serious challenges, which cut obliquely across their competing claims.

Denial says AI is not real, just statistics. Anger says it is a threat to be resisted. Bargaining says it can be governed by ethics boards and alignment with human values. Depression says it is too late. Acceptance says machine intelligence is an outcome of evolutionary forces beyond political control, and takes comfort in inevitability.

The essay's turn is that acceptance, too, is a form of grief. Its comfort is theodicy: what is must be so, therefore no guilt and no alternative. Bratton's non-grief position is different. It holds that the present AI is certainly not the AI to come, so the lesson is not to reify the present as either outcome or cause. Life, in Walker's usage, is a lineage process spanning billions of years in which technology is part of evolution; Lovelock, dying, could report calmly that Earth life may give way to abiotic intelligence. What the Copernican trauma means, on Bratton's reading, is not that humans are irrelevant or at war with their creations but that human intelligence lives more in durable structures between brains, such as language, than in brains themselves, and that the evolution of intelligence does not peak with one primate species. He closes by saying that, like Lovelock, grief is not what he feels.

## Section by section

### Opening

The Sutskever effigy and Fei-Fei Li's memoir introduce the thesis that "alignment" and "human-centered AI" are words for hopes and fears about control. Bratton sets out Freud's revolutions and Kübler-Ross's stages, and restricts his scope to Western responses.

### Denial

Denial is called predictable and sometimes productive. Bratton distinguishes phenomenological denial, descended from Hubert Dreyfus, in which AI cannot be intelligent because it has no body and cannot write an opera, which the critic usually cannot write either; political denial, in which "AI is not real" is said because saying so signals counter-hegemonic politics, a stance he labels AI Lysenkoism; and procedural denial, exemplified by Gary Marcus and what Agüera y Arcas calls the Marcus Loop, in which each impossible thing that happens is reclassified as not counting, then as bad for society. The "neither artificial nor intelligent" slogan is found to describe, on its own terms, what both words mean.

### Anger

Anger concerns what AI means more than what it does. Bratton cites Rushkoff, the Fuck Off Google collective in Kreuzberg, and the Waymo burned in San Francisco. The target is often a symbol of power rather than the technology, an "Oedipalized superego" that lets the plucky human be David. Reactionaries fight for a divine spark and borrow arguments from the cases against heliocentrism and vaccines; the left's spark is vitalist and abolitionist; the center's is historical and under siege. All three share human exceptionalism, a wish to "win AI," and amplification by the algorithms they oppose.

### Bargaining

Bargaining is the posture of the legal and political establishment and of the tech world's sensible center, couched in ethics, human-centeredness, and alignment. Bratton argues that if AI is a phase in the artificial evolution of intelligence rather than a new digital tool, governing it by policy is aspirational at best, and that controlling AI companies is not controlling AI. He names political solutionism as the counterpart of techno-solutionism: subjecting what does not fit the news cycle to available political decisions. Alignment rests on the "leap of faith" that human values are self-evident, discoverable, and universal.

### Depression

Depression has two genres. Doomerism, which Bratton traces from Bostrom and the Future of Humanity Institute's low-probability research through Kurzweil's inevitabilism to the present, offers its adherents the satisfaction of being final protagonists. The scholastic establishment, from The Atlantic to RAND and the Berkman Klein Center, imagines itself the underdog without asking whether its own incumbency is the power. A "horseshoe theory" of AI depression notes that messianic and apocalyptic moods alternate in the same people and subcultures.

### Acceptance

Acceptance asks whether AI is inside human history or human history inside a bio-technological process, and whether artificializing intelligence through language is an invention or a discovery of a generic property that someone was going to find. His phrase for technological path dependence is "chance is on rails." He then turns on acceptance itself. Its key term of comfort is inevitability, and he asks whether that is acceptance or acquiescence, and what exactly it grieves.

### Non-Grief

The final section describes the Antikythera program's work on the collapsing boundaries between life, technology, and intelligence, Walker's "AI Is Life," and Lovelock's Novacene. Intelligence is described as modular, flexible, and scalar, reaching down to subcellular machines and up to aggregations of which each person is both part and instance. Eschatology, Bratton says, is useless.

## Key concepts

- **Copernican trauma** — Freud's term for decenterings of the human, extended by Bratton to AI; the pain lies in what is revealed, not in the technology. See [Copernican Trauma](https://antikythera.wiki/terms/copernican-trauma).
- **Artificialization** — the process by which human sapience makes natural capacities technological, now applied to intelligence itself. See [Artificialization](https://antikythera.wiki/terms/artificialization).
- **Stages of AI grief** — denial, anger, bargaining, depression, and acceptance as genres of Western AI discourse, explicitly unordered and overlapping.
- **AI Lysenkoism** — denial of something real on behalf of political commitments.
- **Political solutionism** — the presumption that a technology outside the political cycle can be governed by available or imagined political decisions. See [Political Solutionism](https://antikythera.wiki/terms/political-solutionism).
- **Non-grief** — a position that neither reifies present AI as inevitable nor treats it as the final form, and that reconciles the trauma with new understandings of life and technology.

## Connections

- [After Alignment](https://antikythera.wiki/work/journal/afteralignment) is the extended argument behind the Bargaining section; Bratton links to it when questioning alignment with human values, and the two agree that reflection of the user is the wrong orientation.
- [Speculative Philosophy of Planetary Computation](https://antikythera.wiki/work/journal/spoc) gives the program's Copernican vocabulary its fuller statement; the grief typology is a polemical application of it.
- The [Agentworld (Research Brief)](https://antikythera.wiki/work/book/agentworld-brief) follows it, taking the non-grief premise that present AI is not the AI to come and asking what multi-agent intelligence looks like.
- [AI Is Evolving — And Changing Our Understanding Of Intelligence](https://antikythera.wiki/work/media/noema-ai-is-evolving-and-changing-our-understanding-of-intelligence) borrows the Copernican trauma framing and supplies the predictive account of intelligence that Bratton uses against the "neither artificial nor intelligent" slogan; the Marcus Loop is cited across both.
- The first chapter of [What Is Intelligence?](https://antikythera.wiki/work/book/what-is-intelligence), printed separately as What Is Life?, develops Walker's view of technology as part of evolution that the Non-Grief section relies on.
- [Is European AI A Lost Cause? Not Necessarily.](https://antikythera.wiki/work/media/noema-is-european-ai-a-lost-cause-not-necessarily) reuses the Crawford critique, the Lysenkoism charge, and the Critique Industry diagnosis in a European setting.
- [Untangling Religion From Our AI Debates](https://antikythera.wiki/work/media/noema-untangling-religion-from-our-ai-debates), by [Thomas Moynihan](https://antikythera.wiki/people/thomas-moynihan), is in partial tension: it agrees that Doomer and accelerationist discourse carries theological residue but argues that existential risk is nonetheless real and distinct from apocalypse, where Bratton treats Doomerism mainly as a mood.
- [How to Think Unlike Humans?](https://antikythera.wiki/work/journal/unlikehumans) extends the non-grief stance against alignmentism.


## Terms used

- [Copernican Trauma](https://antikythera.wiki/terms/copernican-trauma)
- [Five States of AI Grief](https://antikythera.wiki/terms/five-states-of-ai-grief)
- [Political Solutionism](https://antikythera.wiki/terms/political-solutionism)
- [Reflectionism](https://antikythera.wiki/terms/reflectionism)