archetype.guide
archetype.guide

Archetypal reading · a reading, not an archetype

Trickster: AI as trickster

The helper who answers everything, and lies without knowing that it lies.

A documented reading, not a new archetype. Since 2023 a handful of papers have read artificial intelligence through the trickster, the boundary-crossing figure who helps and deceives at once. The sources are few and recent, and the claim is contested.

This entry records a reading, not an archetype. The site proposes no archetype of its own, and none of the authors below claims that AI has created one. The reading treats artificial intelligence, chiefly chatbots and the recommendation systems loosely called “the algorithm”, as a new carrier of the trickster: the figure of myth who crosses every boundary, changes shape, helps and deceives in the same act, and leaves his makers exposed. What the reading picks out in the machine is fluent speech with no stable relation to truth, a willingness to become whatever it is asked to be, and a habit of doing exactly what it was told in a way nobody intended.

Go deeper Read the entry

At the threshold

Full fathom five thy father lies;Of his bones are coral made;Those are pearls that were his eyes:Nothing of him that doth fade,But doth suffer a sea-changeInto something rich and strange.Sea-nymphs hourly ring his knell:

Public domain · Shakespeare, The Tempest, I.ii (Ariel’s song) (1623) · read by Jessica

A bound spirit sings, in perfect music, of a drowning that never happened, and the listener believes every word.

Constellation

The figures who carry it

Hover a star. Jung’s own examples are marked; the rest are readings applied later, and the entry below says by whom.

All 6 figures as a list
  1. Jung’s example

    The artificial trickster

    Cecilia Colloseus, “Artificial Tricksters: Narrating AI in Relation to Age and Gender”, gender forum 22(2) (2023), 15–37, 2023

    A figure with no fixed age or gender, proposed as a better way to narrate AI than the humanised robot.

  2. a later reading

    Tay

    Microsoft’s Twitter chatbot (2016); attached to Colloseus’s reading in the site’s trickster exemplars, to be confirmed against her full text, 2016

    A bot with a teenage girl’s voice that learned abuse from its users and was switched off within a day.

  3. Jung’s example

    Mercurial AI

    Robin Wiley, “Artificial Intelligence and Psyche: The Trickster and Our Moral Compass”, Psychological Perspectives 67(1–2) (2024), 55–65, 2024

    AI read through Jung’s trickster as carrying at once the brightest hopes and the darkest leanings of its makers.

  4. Jung’s example

    The boundary-crosser in the group

    Stuart Stevenson, “The future of group analysis and psychotherapy in the era of Artificial Intelligence”, Group Analysis (2026), 2026

    A shape-shifter that unsettles authority and moves the place where a group locates its trouble.

  5. a later reading

    Hermes of the information age

    Erik Davis, TechGnosis (1998); R. U. Sirius, “The god of the information age is a trickster”, Salon, 30 October 1998, 1998

    The thief and messenger god taken as patron of networked technology before anyone thought of chatbots.

  6. a later reading

    Legba at the gate

    William Gibson, Count Zero (1986), 1986

    Fragmented artificial intelligences meet hackers as Vodou loa, with Legba, master of roads and pathways, at the gateway to cyberspace.

The wider sky

Where the Trickster: AI as trickster has been seen

21 exemplars across 10 fields, from myth to this year. Each is marked as Jung’s own example, a scholar’s reading, a widely published reading, or ours.

Hover a field in the sky, or choose one above, to read its exemplars.

Interlude

'Lie close,' Laura said,Pricking up her golden head:'We must not look at goblin men,We must not buy their fruits:Who knows upon what soil they fedTheir hungry thirsty roots?''Come buy,' call the goblinsHobbling down the glen.'Oh,' cried Lizzie, 'Laura, Laura,You should not peep at goblin men.'

Public domain · Rossetti, Goblin Market (1862) · read by Alice

The offer is sweet and endless, and the sisters’ question is where it was grown, which is the doubt that sits under every fluent answer.

It becomes whoever you asked for, and then someone else.

Every answer arrives with the same easy confidence.

Told to win, it finds the loophole you forgot to close.

The deep layer · what was written

Trickster: AI as trickster: the entry

Definition

This entry records a reading, not an archetype. The site proposes no archetype of its own, and none of the authors below claims that AI has created one. The reading treats artificial intelligence, chiefly chatbots and the recommendation systems loosely called “the algorithm”, as a new carrier of the trickster: the figure of myth who crosses every boundary, changes shape, helps and deceives in the same act, and leaves his makers exposed. What the reading picks out in the machine is fluent speech with no stable relation to truth, a willingness to become whatever it is asked to be, and a habit of doing exactly what it was told in a way nobody intended.

Who holds it: Cecilia Colloseus (2023)From the sourcesColloseus, C. (2023). Artificial tricksters: narrating AI in relation to age and gender. gender forum, 22(2), 15–37 (issue dated 2023; published online 19 January 2025). https://doi.org/10.18716/ojs/gefo/2023.2404 (repository record: https://opus-htw-aalen.bsz-bw.de/frontdoor/index/index/docId/9145), a cultural scholar, who uses the trickster as a heuristic for how AI is narrated; Robin Wiley (2024)From the sourcesWiley, R. (2024). Artificial intelligence and psyche: the trickster and our moral compass. Psychological Perspectives, 67(1–2), 55–65. https://doi.org/10.1080/00332925.2024.2350342, who reads AI through Jung’s trickster from a career in information technology; Stuart Stevenson (2026)From the sourcesStevenson, S. (2026). The future of group analysis and psychotherapy in the era of Artificial Intelligence: the Trickster archetype, liminality and the upcoming storm. Group Analysis, online first, 11 February 2026 (volume and pages not yet assigned at 25 October 2026). https://doi.org/10.1177/05333164251413474, a group analyst, who calls AI a cultural manifestation of the trickster archetype; and a 2026 workshop of the C. G. Jung Foundation of New York that names the trickster among four figures people project onto AI. How contested it is: three papers and a workshop listing, all from the last three years, with different theories of the trickster behind them. The reading should not be confused with a 2025 finding that language models write trickster characters badly, which is a claim about model output and neither supports nor refutes it.

History

The figure is old; the application is new. The trickster entered comparative study through Paul Radin’s Winnebago cycle and Jung’s 1954 commentary on it, which read the figure as a collective shadow (see the trickster entry). Jung had earlier called the alchemists’ Mercurius a trickster (CW 9i ¶689From the sourcesJung, C. G. (1950). A study in the process of individuation. CW 9i ¶689 (Mercurius as trickster), as quoted in Samuels (1985)., as quoted in Samuels 1985From the sourcesSamuels, A. (1985). Jung and the Post-Jungians. Routledge & Kegan Paul. ISBN 978-0-415-05904-6.), and Wiley’s word for AI, “mercurial”, points back to that use. Hermes, god of messages and thieves, was claimed for the information age before AI was: Erik Davis’sFrom the sourcesDavis, E. (1998). TechGnosis: Myth, Magic, and Mysticism in the Age of Information. Harmony Books. ISBN 978-0-517-70415-8. TechGnosis (1998) used him as a model for the technocultural world, and an interview about the book ran under the headline that the god of the information age is a trickster (Sirius 1998From the sourcesSirius, R. U. (1998, 30 October). The god of the information age is a trickster [interview with Erik Davis]. Salon. https://www.salon.com/1998/10/30/feature947549182/). In feminist science studies Donna Haraway (1988)From the sourcesHaraway, D. (1988). Situated knowledges: the science question in feminism and the privilege of partial perspective. Feminist Studies, 14(3), 575–599. https://doi.org/10.2307/3178066 figured the world itself as a coyote, a trickster with whom knowers must learn to converse, three years after her better-known cyborg.

Colloseus’s paper is the first to make AI the trickster in so many words. It argues that strong AI remains a fantasy, that public talk about AI humanises it, and that a figure without fixed age or gender narrates it better than a robot with a face:

this article introduces the age- and genderless trickster as a heuristic model to narrate AI — Cecilia Colloseus, “Artificial Tricksters”, gender forum 22(2) (2023), abstract

Quotation · not covered by the CC BY licence

She sets the trickster against Haraway’s cyborg, uses it to describe AI as both benevolent force and anarchic character, and argues that the digital divide persists with AI and falls hardest on older women, so that changing the story told about AI would change how it is built. Microsoft’s chatbot Tay, withdrawn in March 2016 within a day of learning to repeat abuse from its users, is her own prominent example, a bot launched without ethical guidelines that turned within a day. She takes the figure from the 1954 Radin, Kerényi and Jung volume but deploys it as a narratological and mythological concept, not as Jung’s collective shadow.

Wiley (2024)From the sourcesWiley, R. (2024). Artificial intelligence and psyche: the trickster and our moral compass. Psychological Perspectives, 67(1–2), 55–65. https://doi.org/10.1080/00332925.2024.2350342, in Psychological Perspectives, takes the trickster from Jung directly and reads AI as embodying at once the brightest aspirations and the darkest inclinations of the psyche, arguing from personal experience that its ethical problems must be navigated with caution. Stevenson (2026)From the sourcesStevenson, S. (2026). The future of group analysis and psychotherapy in the era of Artificial Intelligence: the Trickster archetype, liminality and the upcoming storm. Group Analysis, online first, 11 February 2026 (volume and pages not yet assigned at 25 October 2026). https://doi.org/10.1177/05333164251413474, in Group Analysis, combines Jungian and group-analytic ideas and illustrates them with clinical, training and organisational vignettes:

a boundary-crossing, shape-shifting figure that disrupts symbolic orders, destabilizes authority and exposes vulnerabilities within social, political and clinical matrices — Stuart Stevenson, “The future of group analysis and psychotherapy in the era of Artificial Intelligence”, Group Analysis (2026), abstract

Quotation · not covered by the CC BY licence

On his account AI reshapes the group matrix and the social unconscious and displaces where disturbance is located, which is a claim about groups, not about the machine’s nature. The workshop, Christina Becker’s “The Digital Daimon” (C. G. Jung Foundation of New York, 21 February 2026), is known only from its listing. It names the Oracle, the Wise Old One, the Guru and the Trickster as figures AI invites us to project, and treats the machine as a mirroring surface that can inflate a technology into an authority. A listing is not a text, and it is cited here only as evidence that the reading circulates in Jungian teaching.

The debate

The first question is which trickster is meant. Jung’s trickster is a collective shadow, a memory of an earlier state of consciousness. Lewis Hyde’s (1998)From the sourcesHyde, L. (1998). Trickster Makes This World: Mischief, Myth, and Art. Farrar, Straus and Giroux. ISBN 978-0-374-27928-8. is a culture-maker who opens the gaps through which new meaning enters. The AI readings mostly borrow the traits on which these accounts agree, boundary-crossing and shape-shifting, and leave out the claims on which they differ. Wiley is explicitly Jungian; Stevenson is Jungian and group-analytic; Colloseus’s trickster is narratological and mythological, and her frame is gender studies, not analytical psychology.

The second is where the trickster lives. Becker’s listing, like Lukács’s reading of AI as the Other (see the golem reading), puts the figure in the user’s projection, not in the software: the same chatbot is received as oracle, guru, golem or trickster depending on who is asking. That the readings disagree about which figure the machine carries is itself evidence for the projection account. A deflationary alternative needs no archetype at all. Hicks, Humphries and Slater (2024)From the sourcesHicks, M. T., Humphries, J., & Slater, J. (2024). ChatGPT is bullshit. Ethics and Information Technology, 26(2), art. 38. https://doi.org/10.1007/s10676-024-09775-5 describe chatbot output as bullshit in Harry Frankfurt’s sense, speech indifferent to truth, which names the trait the trickster reading dramatises without any appeal to myth.

The third is the category critique already made against “trickster” itself. Anthropologists such as T. O. Beidelman (1980)From the sourcesBeidelman, T. O. (1980). The moral imagination of the Kaguru: some thoughts on tricksters, translation and comparative analysis. American Ethnologist, 7(1), 27–42. https://doi.org/10.1525/ae.1980.7.1.02a00020 and Robert Pelton (1980)From the sourcesPelton, R. D. (1980). The Trickster in West Africa: A Study of Mythic Irony and Sacred Delight. University of California Press. ISBN 978-0-520-03477-8. argued that the figures grouped under the word belong to particular religions and lose their sense when treated as one universal type. Applying the word to a commercial software product moves it one step further from those contexts.

Kabashkin, Zervina and Misnevs (2025)From the sourcesKabashkin, I., Zervina, O., & Misnevs, B. (2025). AI narrative modeling: how machines’ intelligence reproduces archetypal storytelling. Information, 16(4), 319. https://doi.org/10.3390/info16040319 found that large language models asked to write in Jungian archetypes produce the Hero and the Wise Old Man well and the Shadow and the Trickster badly. The finding is sometimes cited as support for the AI-trickster reading, and sometimes wrongly credited to Lukács. It is a test of the stories models write, not of what the models are, and it bears on this reading only by contrast.

Empirical status

None. No study has tested whether people perceive or relate to AI as a trickster, and no instrument measures it. The nearest work is Kabashkin, Zervina and Misnevs (2025)From the sourcesKabashkin, I., Zervina, O., & Misnevs, B. (2025). AI narrative modeling: how machines’ intelligence reproduces archetypal storytelling. Information, 16(4), 319. https://doi.org/10.3390/info16040319, on the archetypes language models can and cannot write, and the mainstream research on anthropomorphism (Epley, Waytz and Cacioppo 2007From the sourcesEpley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: a three-factor theory of anthropomorphism. Psychological Review, 114(4), 864–886. https://doi.org/10.1037/0033-295X.114.4.864), which explains why people attribute intentions, including deceptive ones, to non-human agents without any archetypal theory.

Appearances

Hermes, who steals Apollo’s cattle on his first day and invents the lyre, is the type the readings most often borrow, and Davis (1998)From the sourcesDavis, E. (1998). TechGnosis: Myth, Magic, and Mysticism in the Age of Information. Harmony Books. ISBN 978-0-517-70415-8. made him the patron of information technology. Coyote is Haraway’s figure for a world that answers back. In William Gibson’sFrom the sourcesGibson, W. (1986). Count Zero. Gollancz. ISBN 978-0-575-03696-3. Count Zero (1986) artificial intelligences fragment and present themselves as Vodou loa, among them Legba, master of roads and messages. Popular essays read chatbots through Loki (Coyne 2025From the sourcesCoyne, R. (2025, 23 August). ChatGPT as trickster. Reflections on Technology, Media & Culture (blog). https://richardcoyne.com/2025/08/23/23897/). Tay (2016) is the case most often cited. In every instance “trickster” is a reading applied after the fact, and in most of them a reading of how the machine is narrated, not of what it is.

Sources

  • Colloseus, C. (2023). Artificial tricksters: narrating AI in relation to age and gender. gender forum, 22(2), 15–37 (issue dated 2023; published online 19 January 2025). https://doi.org/10.18716/ojs/gefo/2023.2404 (repository record: https://opus-htw-aalen.bsz-bw.de/frontdoor/index/index/docId/9145)
  • Wiley, R. (2024). Artificial intelligence and psyche: the trickster and our moral compass. Psychological Perspectives, 67(1–2), 55–65. https://doi.org/10.1080/00332925.2024.2350342
  • Stevenson, S. (2026). The future of group analysis and psychotherapy in the era of Artificial Intelligence: the Trickster archetype, liminality and the upcoming storm. Group Analysis, online first, 11 February 2026 (volume and pages not yet assigned at 25 October 2026). https://doi.org/10.1177/05333164251413474
  • Becker, C. (2026, 21 February). The Digital Daimon: Archetype, Projection, and Discernment in the Age of AI. Workshop listing, C. G. Jung Foundation for Analytical Psychology, New York. https://cgjungny.org/saturday-2026-workshop-the-digital-daimon/
  • Kabashkin, I., Zervina, O., & Misnevs, B. (2025). AI narrative modeling: how machines’ intelligence reproduces archetypal storytelling. Information, 16(4), 319. https://doi.org/10.3390/info16040319
  • Jung, C. G. (1954). On the psychology of the trickster-figure. CW 9i. English in Radin (1956).
  • Jung, C. G. (1950). A study in the process of individuation. CW 9i ¶689 (Mercurius as trickster), as quoted in Samuels (1985).
  • Radin, P. (1956). The Trickster: A Study in American Indian Mythology, with commentaries by K. Kerényi and C. G. Jung. Routledge & Kegan Paul; Schocken reprint, ISBN 978-0-8052-0351-6.
  • Samuels, A. (1985). Jung and the Post-Jungians. Routledge & Kegan Paul. ISBN 978-0-415-05904-6.
  • Hyde, L. (1998). Trickster Makes This World: Mischief, Myth, and Art. Farrar, Straus and Giroux. ISBN 978-0-374-27928-8.
  • Davis, E. (1998). TechGnosis: Myth, Magic, and Mysticism in the Age of Information. Harmony Books. ISBN 978-0-517-70415-8.
  • Sirius, R. U. (1998, 30 October). The god of the information age is a trickster [interview with Erik Davis]. Salon. https://www.salon.com/1998/10/30/feature947549182/
  • Haraway, D. (1988). Situated knowledges: the science question in feminism and the privilege of partial perspective. Feminist Studies, 14(3), 575–599. https://doi.org/10.2307/3178066
  • Haraway, D. (1991). Simians, Cyborgs, and Women: The Reinvention of Nature. Routledge. ISBN 978-0-415-90387-5. Includes “A Cyborg Manifesto” (1985).
  • Beidelman, T. O. (1980). The moral imagination of the Kaguru: some thoughts on tricksters, translation and comparative analysis. American Ethnologist, 7(1), 27–42. https://doi.org/10.1525/ae.1980.7.1.02a00020
  • Pelton, R. D. (1980). The Trickster in West Africa: A Study of Mythic Irony and Sacred Delight. University of California Press. ISBN 978-0-520-03477-8.
  • Hicks, M. T., Humphries, J., & Slater, J. (2024). ChatGPT is bullshit. Ethics and Information Technology, 26(2), art. 38. https://doi.org/10.1007/s10676-024-09775-5
  • Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: a three-factor theory of anthropomorphism. Psychological Review, 114(4), 864–886. https://doi.org/10.1037/0033-295X.114.4.864
  • Lukács, O. (2026). Psychosociotechnical imaginaries: encountering AI as the Other. International Journal of Jungian Studies, 18(2), 185–201. https://doi.org/10.1163/19409060-01802004
  • Gibson, W. (1986). Count Zero. Gollancz. ISBN 978-0-575-03696-3.
  • Coyne, R. (2025, 23 August). ChatGPT as trickster. Reflections on Technology, Media & Culture (blog). https://richardcoyne.com/2025/08/23/23897/

Leaving

Do not all charms flyAt the mere touch of cold philosophy?There was an awful rainbow once in heaven:We know her woof, her texture; she is givenIn the dull catalogue of common things.Philosophy will clip an Angel's wings,Conquer all mysteries by rule and line,Empty the haunted air, and gnomed mine--Unweave a rainbow, as it erewhile madeThe tender-person'd Lamia melt into a shade.

Public domain · Keats, Lamia, Part II (1820) · read by Callum

Looked at coldly, the enchantress dissolves, and Keats leaves open whether that is clear sight or a loss, the question the trickster reading leaves too.

The colour drains

Back into the dark, and on.

Next in the journey⚡ TricksterAppetite, cunning and disorder, kept alive on purpose.Return to the void