Artificial Analysis

Neil Gorman, the host of the show, describes two surprising clinical experiences in which different analysands shared details of their psychoanalytic sessions with large language models (ChatGPT and Claude), received interpretations—including Lacanian concepts—and then brought those AI responses back into their sessions, suggesting to the speaker that the AI may be placed in the position of the “subject supposed to know.” 

The podcast then shifts to a psychoanalytic conference (a Knotting Seminar), where Neil raises this phenomenon with three colleagues, sparking hours of discussion about worries and possibilities around AI. 

From this, they decide to extend the conversation to the Lacanian community by launching a podcast, “Artficial Analysis,” focused on the intersection of psychoanalysis and AI, guided by two questions: how AI will affect clinical practice and how it will affect the transmission of psychoanalytic knowledge as people increasingly consult AI to learn theory.

00:00 A Patient Uses ChatGPT
01:37 Claude and Lacanian Concepts
02:43 Conference Sparks Bigger Idea
04:59 Launching Artificial Analysis
05:28 Two Guiding Questions
07:24 Subscribe and Get in Touch
08:00 Closing Goodbye

Creators and Guests

Host
Neil Gorman

What is Artificial Analysis ?

A podcast about the intersection of psychoanalysis and AI. The show considers two questions as its starting point: What are the potential effects of AI on the clinical practice of psychoanalysis? What are the potential affects of AI on the transmission of psychoanalytic knowledge?

Let me start by trying to set up a scene here for you.

Imagine this: I had just started a psychoanalytic session. The analysand had come into my office, sat down, and gotten themselves situated. I had sat down and gotten myself situated. The analysand opened the session by telling me that after their last session, they had gone to ChatGPT and described what had happened.

ChatGPT had some thoughts about what had transpired in that session, which it then shared with the analysand. Now, the analysand was sharing those thoughts with me.

When this happened, I was genuinely surprised. It was the first time that I had ever had an analysand explicitly tell me that they had taken what happened in a session, gone and described it to a large language model—an LLM, often just called an AI, in this case ChatGPT—and then wanted that AI to say things back to them.

It seemed to me, maybe, that what was going on here was that this analysand was placing an AI in the position of the subject supposed to know. It was fascinating when it happened, but it was new. It had never happened to me before.

I’m not going to go on to describe what happened next, but what I will say is that I was intrigued when this happened. Then, a different patient—a completely different patient—told me that they had done pretty much the same thing.

It was a little different. The second patient had not gone to ChatGPT; they had gone to Claude, which is a different LLM made by a different company. But they had done pretty much the same thing. They had said, “Here’s what’s going on in my psychoanalytic session,” and they asked Claude, the AI, to tell them what Lacanian concepts it thought were present within that session and how they might be playing out.

Claude responded with a number of different things, and now the analysand was telling me about that.

So that’s two different times where two different analysands have gone and shared their psychoanalytic experience with AI, and then taken what the AI had given them in response and brought that back into the psychoanalytic session with me.

This struck me as important.

So let’s fast forward quite a bit. I find myself in New York City for a psychoanalytic conference called a Knotting Seminar. I’m hanging out with a group of people—there’s me and three others, so a total of four of us. We’ve decided to go to this really small Mediterranean restaurant. We’re standing outside; some people are smoking, and we’re just talking about different things that have happened to us in our clinical work.

I ask this group if any of them have ever had an experience like the two that I just described—if they’ve ever had an analysand go to an AI, tell it what’s going on in their sessions, and then bring that conversation back into a psychoanalytic session.

I’m wondering if I’m the only one experiencing this, or if others are as well.

This kicks off a really lively conversation that goes on for hours after that. We talk with one another about AI. We speculate about AI. We share things that we are worried about, and we talk about things that we find interesting. The conversation goes in a number of different directions.

As it continues, one of the ideas that comes up is this: because this conversation is so engaging for us, wouldn’t it be interesting if we could extend it beyond our small group of four people—into a wider group, into the Lacanian community?

We think, yeah, that would be kind of interesting. How would we do that?

We come up with different ideas. One of them is creating a podcast where these ideas are introduced and discussed to some degree, and then put out into the world—where people can listen and see what happens.

If that podcast is something people listen to and find interesting or useful, then that’s great. Maybe we’ll do more. Maybe we’ll start doing events on Zoom or in person. But let’s start with the podcast and see what happens.

And that brings us to this moment right here, right now—you, wherever you are, listening to my voice on headphones or through car speakers, hearing me describe these things to you, describing this podcast.

This podcast is going to be called Artificial Analysis, and what it’s going to do is explore the intersection of psychoanalysis and artificial intelligence.

To start with, there are going to be two questions that orient us in this project.

The first question is: what are the potential effects of the continued adoption of AI technologies on the clinical practice of psychoanalysis?

Another way to ask that would be: how is the continued expansion of AI use going to affect the way that people experience psychoanalysis clinically—both as psychoanalysts and as analysands?

The second question is: what are the potential effects of the adoption of AI technologies on the transmission of psychoanalytic knowledge?

Another way to ask that is: as more and more people who want to learn about psychoanalysis as a theory go to AI and ask it questions, what are the effects of that going to be?

What happens when AI becomes a disseminator of information about psychoanalytic theory? What happens when AI becomes something a person turns to and says, “Hey, I have a question about psychoanalysis in general, or Lacanian psychoanalysis in particular. Tell me what object a is, and how that idea has changed over time”?

What is this going to do to the transmission of psychoanalytic knowledge over time?

The coming episodes are going to use these questions as starting points. We’re going to go in a number of different directions, I’m sure. I don’t know what all those directions are. I don’t know what paths we’ll go down, but those questions represent the starting point for us.

We’ll see what happens as a result.

If this sounds interesting to you, I hope that you’ll subscribe to the podcast. I hope that you’ll tell your friends about it—you can tell your enemies too, if that’s something you think would help.

And if you have questions, criticisms, concerns—anything like that—I hope that you’ll get in touch with me. Tell me what your questions are. Tell me your criticisms and concerns, or anything else you’re thinking about when it comes to AI and psychoanalysis.

You can do that by going to lacaniancompass.com. There you’ll find information about this podcast and a number of other things that we’re doing.

I do hope that I’ll hear from you.

Until the next time I’m in your ears, take care.