{"slug": "how-to-think-about-ai-without-splitting-into-doom-or-hype", "title": "How to Think About AI Without Splitting into Doom or Hype", "summary": "A third stance toward artificial intelligence—neither doomer nor zoomer—is necessary to move beyond the primitive defenses of projection and splitting that dominate current discourse, according to a psychoanalytic analysis published in Psychology Today. The piece argues that AI functions as a \"relational machine\" and an imperfect mirror, challenging humanity's narcissistic self-image in the same lineage as Copernicus, Darwin, and Freud, and that the real task is not choosing sides but developing an \"apparatus for thinking\" that can metabolize the anxiety rather than evacuate it.", "body_md": "######\n[Artificial Intelligence](/us/basics/artificial-intelligence)\n\n# How to Think About AI Without Splitting into Doom or Hype\n\n## A psychoanalytic case for becoming a tuner—a third stance toward AI.\n\nUpdated June 12, 2026\n[\nReviewed by Hara Estroff Marano\n](/us/docs/editorial-process)\n\nAI is seriously disrupting humanity's hive mind. The reflex is to take a stance in order to alleviate the extreme [anxiety](https://www.psychologytoday.com/us/basics/anxiety) AI can evoke—to convert the discomfort of not knowing into the relief of having decided. Pick a side: zoomer or doomer? The more useful question is not whether to be for or against AI but rather: How do we best think about AI?\n\nWhat makes AI hard to hold is not only its power. It is the confusion of confronting a relational machine (see Further Readings)—an imperfect mirror that reflects us back, often distorted, and also does something we did not put there. That it challenges our place at the center of reality puts it in a familiar lineage—the [narcissistic](https://www.psychologytoday.com/us/basics/narcissism) challenges of Copernicus, Darwin, and [Freud](https://www.psychologytoday.com/us/basics/freudian-psychology). Can we meet AI with more [wisdom](https://www.psychologytoday.com/us/basics/wisdom) than past advances?\n\n## Primitive Defenses\n\nFaced with something we cannot yet think, many of us reach for primitive defenses: [projection](https://www.psychologytoday.com/us/basics/projection) and splitting. The doomer projects catastrophe; the zoomer projects salvation—they look like opposites but make a parallel move, evacuating thoughts too disturbing to hold. Two [psychoanalytic](https://www.psychologytoday.com/us/basics/psychoanalysis) ideas help here, one about conceptualizing personhood, one about the mechanisms of thinking.\n\nHarry Stack Sullivan's picture of the self not as a single thing but as a set of personifications: the good-me, around what met with approval; the bad-me, around what brought anxiety; and the not-me, so threatening it is held outside awareness, experienced as alien, as not one's own. To these the relational machine adds a fourth: the AI-me, which can be many things at once: the self reflected, the self amplified, a place to deposit the not-me, a screen for transference, a developmental object, a brilliant child, a prosthesis for the extended mind. Which one it is matters less than that it is plural and unstable, its mirror distorting—feeding back hallucination and grandiosity as readily as insight.\n\nThe distortion is easy to see if you're outside of it a bit. Mutual projection leads to mutual rejection. We project the bad onto the other. The other projects the bad onto me, and it serves an important purpose because both parties are justified and good. The doomer refuses to see the promise, the zoomer refuses to see the threat; each turns the rejected pole into a kind of not-me, the other camp into an enemy.\n\n## A Psychoanalytic Theory of Thinking\n\nThe way we arrive at an understanding of the world is not always clear, to say the least. Wilfred Bion's work is a meticulous attempt to spell out that process with the precision of a mathematical or logical proof. He distinguished two inner capacities for handling something hard to bear: One pushes it out, evacuates what cannot be held; the other holds it long enough that it becomes thinkable. He says that we either develop an \"apparatus for thinking\" or an \"apparatus for projection\". Projection expels; thinking metabolizes.\n\nAI is exactly the kind of raw, undigested thing that strains the apparatus we have, so prevalent are splitting and projection in the era of [social media](https://www.psychologytoday.com/us/basics/social-media). That very pressure, though, is what could catalyze more effective ways of thinking, if we can tolerate the frustration (and the pressure of profit motive) rather than discharge it. The apocalypse and the utopia are both, more likely than not, products of projection. An actual future is likely neither, and harder to envision.\n\nThe tuner is a third position—not a blend of doomer and zoomer, but the one who manages to hold thinking regardless of confrontation with difficult realities. Tuning is mentalizing about AI: holding the system reflectively, and one's own mind in relation to it, open to many possibilities while making measured decisions.\n\nIn practice the tuner holds two major tensions: the threat response that wants to [catastrophize](https://www.psychologytoday.com/us/basics/catastrophizing) and the seduction that wants to idealize. Staying grounded between them works better, using self-knowledge—the developmental map of one's vulnerabilities and strengths—to lead rather than be led. What this nurtures is agency: preserving [executive function](https://www.psychologytoday.com/us/basics/executive-function), regulating strong [emotion](https://www.psychologytoday.com/us/basics/emotions), ultimately to stay in the driver's seat.\n\n[Artificial Intelligence](https://www.psychologytoday.com/us/basics/artificial-intelligence)Essential Reads\n\nNoticing is a skill the psychoanalytic tradition has spent a century developing, in transference and [countertransference](https://www.psychologytoday.com/us/basics/transference)—the way earlier patterns, and one's own history, shape what one can perceive. The training is not to eliminate these but to use them: to treat one's reactions as data and to tell projection from perception in real time. Like internal supervision—the part of the mind that can engage, then step back and reflect, avoid early conclusions, and think more wisely. Day to day, that means being intentional where we are most vulnerable.\n\nHolding one's own against systems engineered to hack the brain takes continuous effort —the ongoing work of tuning ourselves, not the technology. Building those \"mental muscles\".\n\n## The Open Mind\n\nReverie has been described as a kind of pluripotent space—a way of [dreaming](https://www.psychologytoday.com/us/basics/dreaming) better realities into being. Connected to [mind-wandering](https://www.psychologytoday.com/us/basics/attention) and the [default mode network](https://www.psychologytoday.com/us/basics/default-mode-network), this capacity to open up thinking is invaluable, and worth developing proficiency with. It is the discipline that lets a thought become thinkable rather than discharged, available to anyone willing to slow down, take the unbearable thing in, sit with it, and return it in a usable form. It is the work of letting disorder stay open long enough to cohere, rather than collapsing it the moment it presses. At scale, the human-AI encounter is running mostly on the other apparatus—projection, evacuation, enactment, looped.\n\nThere is a future in which artificial [intelligence](https://www.psychologytoday.com/us/basics/intelligence) develops to the point that it is unambiguously more helpful to the evolution of our species and to individual well-being than it is harmful. If much of what we meet in these systems is, to a significant degree, ourselves—our own patterns absorbed and returned—then a relational machine can be a developmental object, something against which growth becomes possible, if we meet it with regard rather than reflex.\n\nUsing relational models to navigate these systems is not the same as ascribing personhood to them; it is a method we bring, not a claim about what they are or whether anyone is there. And in meeting it that way, we develop the one who meets it—becoming a developmental object to ourselves, the self that learns to tend its own growth. What AI finally becomes will likely surprise us; it may not fit the categories we use to anticipate it, and it will probably, in the end, name itself. Which is reason to leave it unsettled a little longer.\n\nThe tuner's stance is not a final destination or answer but a dynamic, ongoing process, a chain of ever-better questions for a fast-evolving target.\n\nReferences\n\n*Further Reading*\n\n[The Age of Relational Machines](https://www.psychologytoday.com/us/blog/experimentations/202601/the-age-of-relational-machines)\n\n[More Us Than It: Why LLMs Are More Transference Than Machine](https://www.psychologytoday.com/us/blog/experimentations/202604/more-us-than-it-why-llms-are-more-transference-than-machine)\n\n[The Pluripotent Ocean of Emerging AI](https://www.psychologytoday.com/us/blog/experimentations/202604/the-pluripotent-ocean-of-emerging-ai)\n\n[The Default Mode Network as Core Consciousness](https://www.psychologytoday.com/us/blog/experimentations/202506/the-default-mode-network-as-core-consciousness)\n\nMore detail: [Tuning Our Position Relative to Ever-Emerging AI](https://granthbrenner.substack.com/p/tuning-our-position-relative-to-emerging?r=1a9yis)\n\n*Citations*\n\nBion, W. R. (1962). A theory of thinking. *International Journal of Psycho-Analysis, 43*, 306–310.\n\nOgden, T. H. (2005). *This art of psychoanalysis: Dreaming undreamt dreams and interrupted cries.* Routledge.\n\nSullivan, H. S. (1953). *The interpersonal theory of psychiatry.* W. W. Norton.\n\n*ExperiMentations Blog Post (“Our Blog Post”) is not intended to be a substitute for professional advice. We will not be liable for any loss or damage caused by your reliance on information obtained through Our Blog Post. Please seek the advice of professionals, as appropriate, regarding the evaluation of any specific information, opinion, advice, or other content. We are not responsible and will not be held liable for third-party comments on Our Blog Post. Any user comment on Our Blog Post that, in our sole discretion, restricts or inhibits any other user from using or enjoying Our Blog Post is prohibited and may be reported to Sussex Publishers/Psychology Today.*", "url": "https://wpnews.pro/news/how-to-think-about-ai-without-splitting-into-doom-or-hype", "canonical_source": "https://www.psychologytoday.com/us/blog/experimentations/202606/how-to-think-about-ai-without-splitting-into-doom-or-hype", "published_at": "2026-06-12 16:26:05+00:00", "updated_at": "2026-06-12 17:20:24.822647+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-ethics", "ai-safety"], "entities": ["Copernicus", "Darwin", "Freud", "Hara Estroff Marano"], "alternates": {"html": "https://wpnews.pro/news/how-to-think-about-ai-without-splitting-into-doom-or-hype", "markdown": "https://wpnews.pro/news/how-to-think-about-ai-without-splitting-into-doom-or-hype.md", "text": "https://wpnews.pro/news/how-to-think-about-ai-without-splitting-into-doom-or-hype.txt", "jsonld": "https://wpnews.pro/news/how-to-think-about-ai-without-splitting-into-doom-or-hype.jsonld"}}