Some Reflections on Intervals, Mediums, and Mediation

By James Curley-Egan

There is an old joke about a turtle who, while crossing the road, is mugged by two snails. When the police show up to ask for a description of his assailants, all he can manage is, “I don’t know, it all happened so fast.”

It’s a joke that has come to mind as I have tried to write intelligibly about Artificial Intelligence (AI); like the poor turtle, who, let’s be honest, should have been able to see snails—I feel defeated, a victim of terrapin temporality; everything seems to be happening too fast. In the relatively small interval that has passed since I began reflecting on this subject, an interval of just a few months, the entire landscape seems to have shifted.[1] I will include what my original thoughts had been (which, like a seismograph, could be read as having unwittingly registered the kind of shift I am evoking), but first I will describe the shift itself.

Just six months ago, only a handful of my patients were sharing their experiences with chatbots. Their reports sounded like what was in the newspaper: the time saved making presentations; the ease with which one could write code or term papers; the sometimes funny, sometimes worrying hallucinations. AI was part of the Umwelt, like the news cycle or the changing seasons or the UN General Assembly—a tool, a gadget, an (external) object.

Since then, it seems to have also taken up residence in the Innenwelt. It is not just the frequency, which is daily, but the way that AI enters the room, which is partly to say the way that AI is brought to bear directly on the work itself: Claude has a novel understanding of one patient’s symptoms; ChatGPT can describe Borderline Personality Disorder in a way that resonates with another patient (why haven’t we discussed “BPD” before?); but what is most impressive and what carries an unmistakable sense of rebuke is the immediacy.[2] It is not merely the breadth or depth of what AI seems to understand but the speed with which that understanding comes across—an implicit disavowal of the ethos of not-knowing-yet, of the abeyance that psychoanalytic work engenders.

The clinician is confronted with an intruder: what is this countertransference? Competitiveness? Self-doubt? Is faith in psychoanalytic process to blame? Is it possible that the machine sees something the clinician doesn’t (or worse, can’t)?

Hence the feeling of shifting ground, which I take as a signal of some primitive anxiety. I had mused with colleague and “AI in My Mind” series editor, Amy Levy, some months ago about the way reflections about AI can take on the trappings of adolescence (which I will only touch on here): like an adolescent, commentaries about AI sometimes seem to be in a rush to declare, once and for all, what Artificial Intelligence is—a compensation, as our work makes clear, for just how terrifying it can be to have that identity shaken. As with adolescence, it appears that a fundamental shift is afoot in the dialectic between inside and outside. It is not quite clear whether a psychiatrist would diagnose that patient as borderline or whether such a diagnosis would be useful to them; but it is clear that ChapGPT is facilitating an enactment—and it is this facilitation, and what it says about (non-artificial) intelligence, that deserves more attention.

Medium

I will return to this question shortly. In the meantime, and with the prospect of enactment still resonating, it feels useful to bring out what I had begun to write a few months ago: some reflections on what it might mean for The Psychoanalytic Quarterly to begin publishing on Substack and, fortuitously, how our presence in this medium might relate to our subject, AI.

In an analytic session, there are some constants: the room, the décor, the chair, the couch, the fee, the hour. In fact, and as is fairly well established, it is the constancy of these elements that allows our work to take place, a fortuitous figure of speech; for it is by means of establishing a space, a scene (which is, after all, how Freud had originally conceived of the unconscious[3]), that these elements hold steady, establishing the frame for what will fill it. Like any good boundary, they help distinguish as clearly as possible what happens inside the session from what happens outside of it, allowing everything within the session to become legible—including the analyst themself—for the ways they might bear the marks of the patient’s unconscious.

To gloss over the particularities of the frame or to leave them unacknowledged, is thus to risk letting that boundary be porous, which—though certainly the clinician’s prerogative—thereby also risks letting all sorts of outside phenomena enter the work, leading to enactment. (Again, as I want to make clear, this is not inherently verboten, but a matter of an analyst’s tolerance.) But it remains no less true that our technique rests upon the steadiness of the frame, which isn’t to say that the frame can’t evolve, but rather that we should pay very close attention to what happens whenever the frame does evolve.[4]

As The Psychoanalytic Quarterly enters this new-ish medium of Substack, and long before we feel held enough for meaning to emerge, we would be right to consider what the medium itself might indicate about the situation of the discourse (to also in tandem consider Marshall McLuhan’s oft-cited phrase: “the medium is the message”). Just as when a patient first enters a consulting room, and whether they choose to say so or not, the room is constructed anew. For what is published here is different than what might appear in a printed journal; it is not printed on a page at all, but is projected, by some wizardry, on a screen; it could be in any color (it could even change color, as it does now when I highlight words and edit them); it could be in any font, size, or orientation; it might be surrounded by text or images or short videos, which in turn might have been placed there by a human or an algorithm (itself designed by a human or by a computer). In fact, by virtue not only of the ostensible subject of this column (AI) but of our particular moment in history, these words might also have been composed by a human or by a computer.[5]

The type of reading one does on a screen or in a web browser has become so normalized that it feels anachronistic to even call attention to it. But it is a part of our orientation to slow down (like a turtle?), to continually pose the question of the present, of “what is happening right now?” And among whatever else might be happening right now, we can take the minimal step in asserting that the discourse has called forth these short entries to be read here, in this relatively new medium. Why it has done so will only be half-understood or deferred; hence the additional significance that has emerged in even so short a span as a few months.[6]

But it does feel apropos in enumerating all of the variables of this new medium, and thus the malleability of its frame, to underscore that our chosen subject is Artificial Intelligence. It is a confluence that we might treat as an association: to raise the question not only of what it might mean for The Psychoanalytic Quarterly to take up AI, but of what it might mean that we are addressing AI in a different medium. It all seems to be happening so fast.

This “happening so fast” is close to the heart of what I think I want to say. It calls to my mind a passage in Walter Benjamin’s “The Storyteller” (which Giorgio Agamben cites in the opening of Infancy and History):

Men returned [from war] grown silent—not richer, but poorer in communicable experience….What ten years later was poured out in the flood of war books was anything but experience that goes from mouth to mouth. And there was nothing remarkable about that. For never has experience been contradicted more thoroughly than strategic experience by tactical warfare, economic experience by inflation, bodily experience by mechanical warfare, moral experience by those in power. A generation that had gone to school on a horse-drawn streetcar now stood under the open sky in a countryside in which nothing remained unchanged but the clouds, and beneath these clouds, in a field of force of destructive torrents and explosions, was the tiny, fragile human body.

To which Agamben adds:

Today, however, we know that the destruction of experience no longer necessitates a catastrophe, and that humdrum daily life in any city will suffice. For modern man’s average day contains virtually nothing that can still be translated into experience. Neither reading the newspaper, with its abundance of news that is irretrievably remote from his life, nor sitting for minutes on end at the wheel of his car in a traffic jam. Neither the journey through the nether world of the subway, nor the demonstration that suddenly blocks the street. Neither the cloud of tear gas slowly dispersing between the buildings of the city centre [sic], nor the rapid blasts of gunfire from who knows where; nor queuing up at a business counter, nor visiting the Land of Cockayne at the supermarket, nor those eternal moments of dumb promiscuity among strangers in lifts and buses. Modern man makes his way home in the evening wearied by a jumble of events, but however entertaining or tedious, unusual or commonplace, harrowing or pleasurable they are, none of them will have become experience.[7]

These are passages that warrant (and reward) being read more than once. There is a sense in which Agamben is describing what happens when a frame explodes, as it did repeatedly over the course of the twentieth century. It isn’t part of his vocabulary, nor really his thesis, to reflect on the way that modern life entails an attack on linking, but the regularity and repetitiveness of disruption,[8] and thus the imbrication of a schizoid relationship to the world (with Agamben, one should be careful not to call this relationship “experience”), is certainly a part of the shape that the discourse has taken. So while it might feel anachronistic, or even wistful, to bemoan the type of reading one can do with a printed, bound, text, it also feels like it would take a kind of superhuman (or superturtle) intelligence, if not an artificial one, to be able to account for the myriad variables that frame this one.

To put it another way, perhaps there is something about doing reflection on A.I. that inheres to this digital, short-form, rapid turnaround, read-it-in-one-sitting medium. Here is the place where the dialectic of this essay begins to resolve. The message of this medium appears to have something to do with intervals: the intervals of my writing process; the interval between clinical sessions (which AI can fill or negate); the interval in which the snails were able to mug the turtle; the unprecedently short interval, with Benjamin and Agamben, between events in modern life; the interval between “now” and “before,” which seems to be increasing in both frequency and absoluteness. What this medium, the Quarterly on Substack, appears to be requesting, even demanding, is that we pay attention to these intervals—that we do what we can to account for what intervals are and what they do, such as, for instance, underwrite the shape of a frame.

What remains to be done, then, and what, happily, clinicians are uniquely trained to do, is the very thing adolescents tend to resist: what does it feel like? This medium—in between patients, or on one’s phone, in one sitting (though also likely interrupted), and having maybe forgotten how the ending of these two thousand words relate to the beginning (feel free to scroll up!)—what is the experience of this medium? It is explicitly about AI, but it is also thereby of AI; what, if we begin to establish a link, does AI feel like in this medium?

I certainly hope to have the opportunity, in the coming months, to deconstruct some of the epiphenomena that accompany our AI present: how AI trades in metalanguage; how it interfaces with Empathy (the bête noire of so many tech evangelists); how we could pose the question of whether A.I. even exists; and so on. But for now, as part of my entree, I want to repeat, maybe in clearer terms, what I have been trying to articulate for the past few minutes.

What it seems to me that this moment of psychoanalytically-considering-AI-on-Substack is asking for, is to not only take account of how we have introjected this new part object, a technical, extra-human part object, but to do so after having already (been) identified with it. Our technique tells us that it might not be quite time to interpret; but merely to exist, to contain. As Michael Balint reminds us:

The analyst…is not so keen on ‘understanding’ everything immediately, and in particular, on ‘organizing’ and changing everything undesirable by his correct interpretations; in fact, he is more tolerant towards the patient’s sufferings and is capable of bearing with them—i.e. of admitting his relative impotence—instead of being at pains to ‘analyse’ them away in order to prove his therapeutic omnipotence.[9]

It feels intent upon us, that is, to accept that this change in medium is a signal that something like projective identification is taking place—for only then can we begin the complicated archaeology of what AI might be splitting off. In the meantime, and though I sense I have already extracted enough from our poor turtle, we would do well to accept the terms of his sensorium; to consider the snail not from the perspective of the police, those agents of executive power, of omnipotence, but from the perspective of our own relative (tortoise) impotence.

---

This article is part “AI in My Mind,” a series about AI edited by Amy Levy.

About the Series:

AI has been in humanity’s mind for a long while. Much like the longing in James Taylor’s nostalgic song, “Carolina in my mind”—written while Taylor was in London missing North Carolina—humanity, in creating AI systems, is expressing a yearning for something that, like memories of a place from one’s past, exists only in the mind. The rich collection of human fantasies about AI, accessible through science fiction stories and films, both buttresses and distorts our current experiences with all too real AI systems. “AI in my mind” is an ongoing series that invites psychoanalytic thinkers to reflect on human desires and goals in creating AI, as well as the effects of real AI systems in our minds. Submissions are invited from psychoanalytic clinicians and scholars who are grappling with AI. During a time when AI is changing humankind, and evoking fear as well as hope, tensions mount and psychoanalytic engagement is challenging to sustain. “AI in my mind” offers us a place to slow down and use psychoanalytic thinking to process, understand, and react to this fast-moving creation.

---

About the Author

James Curley-Egan (PhD, LP, FIPA) is a psychoanalyst in New York City as well as faculty at the Academy for Medicine and the Humanities at the Mount Sinai School of Medicine.

---

About the Editor

Dr. Amy Levy is a clinical psychologist and psychoanalyst. She chairs the American Psychoanalytic Association’s (APsA) President’s Commission on Artificial Intelligence (CAI), serves on the editorial board of The Psychoanalytic Quarterly, and on the subcommittee “Artificial Intelligence” for the International Psychoanalytical Association (IPA) committee, Psychoanalysis and Technology. Dr. Levy teaches and lectures internationally on the intersection of psychoanalysis and artificial intelligence, and is the author of The New Other: Alien Intelligence and the Innovation Drive. She maintains a private practice in Chapel Hill, North Carolina.

---

[1] Articles like this one justify that sense: https://www.forbes.com/sites/lanceeliot/2025/05/14/top-ten-uses-of-ai-puts-therapy-and-companionship-at-the-1-spot/

[2] It lends credence to the thesis of Danielle Knafo and Rocco Lo Bosco’s The Age of Perversion (Routledge 2017).

[3] Or that phrase he borrows from Fechner: the scene of action in dreams; eine andere Schauplatz (an other scene), from SE IV, 48 and V, 536.

[4] The generational shift that the COVID pandemic brought about is lurking here—for instance in the subtle ways that, reflexively, we regularly deny the absolute disruption to the Weltanschauung (Substack belongs to this disruption).

[5] To be clear: I have made no deliberate use of any AI tools while writing or researching this column—but if what I am (attempting) to write here has any meaning, isn’t that irrelevant? This position is just a recovery of similar propositions made more eloquently (and with far fewer digital resources) by thinkers throughout history, though I have in mind Donna Haraway’s “Cyborg Manifesto” from 1985. Cf. Haraway, Donna. Simians, Cyborgs and Women: The Reinvention of Nature (New York; Routledge, 1991), pp.149–181.

[6] What is more: as Cara Maniaci recently wrote for this very series (which only appeared to me in the interval between when I started writing and when I’ve finished it), we might read in this Substack medium a response to a new kind of object—of a kind that produces a fundamental shift in the ground and which (unwittingly?) evokes this kind of change: Media Mother. It is worth remembering that psychoanalysis has from its very origins functioned as a medium—which I mean here in the occult sense i.e., as a medium for making contact with the supernatural. Cf. Hannah Zeavin’s “Freud’s Séance,” in American Imago 75.1 (Spring 2018).

[7] Agamben, Giorgio. Infancy and History: The Destruction of Experience. Trans. Liz Heron. Verso: New York, 1993. 13-14.

[8] I am thinking of the way that this term is glorified in Silicon Valley culture—caricatured, for instance, in Miles Bron (Edward Norton) in the Netflix film Glass Onion: A Knives Out Mystery.

[9] Balin, Michael. The Basic Fault. Tavistock, 1968, 184. I have pursued this quote via Ogden’s 1979 paper “On Projective Identification”, IJP 60, 366.