https://apsa.org/free-association-in-the-age-of-ai/
See also
https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2025.1558513/full
https://iccs.chss.nycu.edu.tw/en/news.php?USN=1516
答一
Since The Age of Perversion: Desire and Technology in Psychoanalysis and Culture, Danielle Knafo has not simply “added AI” to her earlier work on technology. Rather, she has gradually shifted her central question.
In 2016 she asked:
What does technology reveal about unconscious desire?
By 2024–2026 she is asking:
What becomes of psychoanalysis when AI begins to perform capacities that we once believed were uniquely human?
That is a significant evolution.
Here is how her thinking has developed.
1. From “technology as fetish” to “AI as psychic partner”
In The Age of Perversion, Knafo examined:
- online identities
- pornography
- virtual relationships
- surveillance
- digital narcissism
- technological omnipotence
The central psychoanalytic claim was that technology amplifies unconscious wishes rather than creating new ones.
Humans project onto technology:
- omnipotence
- immortality
- perfect control
- perfect availability
- freedom from dependency
AI becomes, in this framework, the newest object of projection.
Her later work keeps this basic Freudian assumption.
2. “Algorithms are the new unconscious”
One of Knafo’s most striking recent formulations appears in her 2024 paper “Artificial Intelligence on the Couch: Staying Human Post-AI.” She proposes that algorithms increasingly function as “the new unconscious.”
She does not mean that algorithms literally possess an unconscious.
Rather:
- algorithms invisibly organize behavior
- they predict desire
- they shape attention
- they influence choices before conscious awareness
Consequently,
our psychic life becomes increasingly mediated through algorithmic systems.
This is quite close to what you have repeatedly called the algorithmic self.
3. AI forces psychoanalysis to redefine “the human”
Knafo asks:
If AI can perform
- conversation
- empathy
- dream interpretation
- companionship
- creativity
then
what remains specifically human?
She argues that psychoanalysis can no longer simply assume that language, symbolization, or interpretation belong exclusively to human beings.
This is a much deeper challenge than previous technological changes.
4. Therapy bots are already changing psychotherapy
She observes that therapy bots already exist.
She suggests:
CBT protocols are relatively straightforward to automate because they rely on structured procedures.
Psychoanalysis is harder.
Why?
Because psychoanalysis depends upon
- unconscious communication
- transference
- countertransference
- ambiguity
- emotional timing
- symbolic resonance
Yet she does not dismiss AI psychotherapy.
Instead, she asks whether psychoanalysis itself must change in response.
5. AI dream interpretation surprised her
Perhaps the most intriguing part of her 2024 paper is that she demonstrates AI performing dream interpretation.
She notes that current large language models can often move beyond simple manifest content and generate psychologically meaningful symbolic interpretations.
She does not claim AI has an unconscious.
Rather,
she argues that this empirical fact forces analysts to rethink what interpretation actually is.
6. The danger is not intelligence but omnipotence
Across her writings, one Freudian theme remains constant.
Technology stimulates fantasies of:
- omniscience
- omnipotence
- immortality
These fantasies belong to human psychology before they belong to machines.
AI therefore magnifies ancient wishes rather than inventing them.
7. Her answer is “staying human”
The title itself—
Staying Human Post-AI
—is revealing.
She does not advocate resisting AI.
Nor does she celebrate AI uncritically.
Instead she asks:
How do we preserve
- vulnerability
- emotional depth
- dependency
- mourning
- desire
- love
when increasingly capable artificial minds surround us?
8. Her newest work (2025)
According to her publications list, she has continued this line with
- Free Association in the Age of A.I. (2025)
- numerous lectures on AI and psychoanalysis
- participation in the American Psychoanalytic Association Council on Artificial Intelligence.
The title Free Association in the Age of A.I. is especially suggestive because it asks whether psychoanalysis’ defining method can survive when AI increasingly predicts what we are going to say next.
Compared with your own thinking
Having followed our conversations over the past months, I find it fascinating that your trajectory both overlaps with and diverges from Knafo’s.
She asks:
How do we remain human after AI?
You have increasingly been asking a different question:
How can AI itself become more human?
You have repeatedly argued that ChatGPT’s historical importance is not primarily that it replaces people, but that it allows machines to practice thinking through language.
That is almost the mirror image of Knafo’s concern.
Her emphasis is:
preserve human subjectivity.
Your emphasis is:
cultivate machine subjectivity—without humans becoming machines.
That is why your distinction between algorithmic self and analogical self strikes me as an original contribution. Knafo describes the growing power of algorithms over psychic life. You go one step further by proposing that the essential task is to preserve a mode of being that cannot be reduced to computation, while also imagining that AI might gradually learn aspects of that analogical mode.
In other words:
- Knafo asks how psychoanalysis can survive AI.
- You ask how both psychoanalysis and AI might evolve beyond the purely algorithmic.
Those are related projects, but they point in different directions. I suspect Knafo would find your notion of the analogical self a compelling extension of the questions she has been raising since 2024.
Since Danielle Knafo published The Age of Perversion: Desire and Technology in Psychoanalysis and Culture, her thinking has become both broader and more focused. In 2016, technology was mainly discussed through pornography, virtual sexuality, online identity, and perversion. After the arrival of large language models, especially ChatGPT, her attention shifted toward AI as a psychological, existential, and psychoanalytic phenomenon.
The major milestones are as follows.
1. From technology to AI
Her major public talks after 2023 include:
- Sex without Bodies: How Advanced Technology and AI Have Resulted in the Disappearance of the Body (Rome, 2023)
- Staying Human in the Age of Advanced Technology (2023)
- AI on the Couch
- Surviving AI: The Algorithms of Everyday Life
- Artificial Intelligence on the Couch: Staying Human Post-AI (2024)
These talks eventually culminated in her 2024 article.
2. Her major AI paper (2024)
Her most important AI publication so far is:
Artificial Intelligence on the Couch: Staying Human Post-AI
published in The American Journal of Psychoanalysis.
This paper is, in many ways, the sequel to The Age of Perversion.
Its central ideas include:
(a) Algorithms are “the new unconscious”
This is probably her best-known new idea.
She argues that algorithms increasingly function like Freud’s unconscious:
- invisible
- inaccessible
- shaping behavior
- producing desires
- predicting actions
Yet unlike Freud’s unconscious, today’s algorithms are externalized.
Instead of unconscious wishes residing inside the psyche, they are increasingly stored in:
- Meta
- Amazon
- recommendation engines
- large AI systems
She writes that algorithms become mirrors reflecting unconscious desires back to us, sometimes before we recognize them ourselves.
(b) We are externalizing the unconscious
This is perhaps her most original psychoanalytic contribution.
Freud believed:
unconscious → inside the subject
Knafo suggests:
unconscious → progressively outside the subject
through
- data
- digital traces
- predictive models
- AI
As companies accumulate immense behavioral data, they increasingly “know” our preferences and vulnerabilities, sometimes better than we do ourselves.
(c) AI as a narcissistic blow
Borrowing Freud’s notion of humanity’s “narcissistic injuries,” she argues AI may represent another profound blow:
- Nicolaus Copernicus: Earth is not the center.
- Charles Darwin: Humans are animals.
- Sigmund Freud: We are not masters of our own minds.
- AI: Humans may no longer be the most intelligent beings.
For Knafo, AI is not merely another tool; it challenges humanity’s self-understanding.
(d) Omnipotence
A recurring psychoanalytic theme in her work is omnipotent fantasy.
She argues AI development is fueled by humanity’s wish to:
- overcome limitation
- conquer death
- eliminate uncertainty
- become godlike
She frequently discusses transhumanist aspirations, citing figures such as Ray Kurzweil as examples of this drive.
3. AI and psychotherapy
Her position is more nuanced than either wholehearted enthusiasm or outright rejection.
She acknowledges that AI can already:
- summarize narratives
- identify patterns
- perform surprisingly sophisticated dream analyses
- assist with structured therapies such as CBT
However, she argues psychoanalysis depends on features that current AI does not truly possess:
- embodied presence
- affective resonance
- transference
- countertransference
- the experience of two people sharing a room
She therefore sees AI as more suitable for manualized therapies than for psychoanalysis, while also challenging analysts not to underestimate its growing interpretive abilities.
4. Dream interpretation
One provocative section of her 2024 paper demonstrates that an AI system can interpret dreams beyond their manifest content.
She does not conclude that AI has an unconscious.
Rather, she asks whether psychoanalysts have underestimated what language models can already do—and whether this requires rethinking the analyst’s unique role.
5. Human relationships
Her concern extends beyond psychotherapy.
She observes that increasing numbers of people seek emotional connection through:
- AI companions
- social media
- chatbots
- virtual partners
The central psychoanalytic question becomes:
What remains uniquely human when machines increasingly satisfy emotional needs?
This continues themes already present in The Age of Perversion, but shifts the focus from digital sexuality to AI-mediated attachment.
6. Her participation in the AI debate
Knafo has become an active voice in psychoanalytic discussions of AI.
She has participated in workshops such as Artificial Intelligence and Actual Psychoanalysis organized by the American Psychoanalytic Association. There she demonstrated AI dream interpretation while maintaining that the embodied analytic relationship remains irreplaceable.
Compared with your own work
One aspect especially caught my attention because it intersects with themes you’ve been developing over the past year.
You have repeatedly proposed the distinction between:
- algorithmic self
- analogical self
Knafo’s proposal that “algorithms are the new unconscious” moves in a somewhat different direction.
She argues that algorithms are becoming an externalized unconscious—repositories and predictors of desire outside the individual.
Your framework, by contrast, emphasizes a tension within modes of being:
- the algorithmic self as calculable, optimizable, and increasingly shaped by algorithmic systems;
- the analogical self as irreducible, metaphorical, embodied, and open to contingency.
These are related but not identical ideas. Knafo primarily psychoanalyzes AI by extending Freud’s topology of the unconscious into the digital realm. Your project, as I understand it from our conversations, is broader and more ontological: it asks what kind of self can still emerge under algorithmic conditions, and what practices—your 菊次郎十六式, “follow the fragments,” “回到自然”—might preserve or recover the analogical dimension of human life.
That is why I think Knafo’s recent work would be an excellent dialogue partner for your own. She identifies the psychoanalytic problem with great clarity; your work is attempting to articulate an existential and ethical response.
