Anthony Elliott is, in my view, one of the most important social theorists for understanding AI—not because he explains how AI works technically, but because he asks what AI is doing to human subjectivity.
That distinguishes him from many philosophers of AI.
- AI researchers ask: How can we build intelligent machines?
- Ethicists ask: How should we regulate AI?
- Elliott asks: What kind of people are we becoming because we live with AI?
This concern grows directly out of his earlier work on psychoanalysis, identity, globalization, and mobility. AI is not a break with his earlier thinking; it is its continuation.
His work over roughly the last decade can be seen as an unfolding sequence.
Stage 1 (2019):
The Culture of AI
This book asks the basic question:
How does AI become part of everyday life?
Elliott argues that AI is no longer confined to laboratories.
It now quietly inhabits
- phones,
- homes,
- hospitals,
- cars,
- workplaces,
- schools,
and increasingly mediates ordinary routines.
His point is not that robots are taking over.
Rather,
AI is becoming
the invisible background of social life.
This is why he calls it
the culture of AI.
Culture changes before politics notices.
Technology becomes ordinary before we realize how much we have changed.
Stage 2 (2022):
Making Sense of AI: Our Algorithmic World
This is, I think, his most important AI book.
Here he introduces what has become his signature concept:
algorithmic modernity.
Algorithmic modernity means that
algorithms are no longer merely tools.
They increasingly organize
- perception,
- work,
- relationships,
- consumption,
- identity,
- politics.
The question therefore becomes
not
“What can AI do?”
but
“What kind of society emerges when algorithms quietly coordinate everyday life?”
He repeatedly emphasizes that AI should be understood as a social transformation, not merely a technological innovation.
Stage 3 (2023):
Algorithmic Intimacy
This is where Elliott becomes especially original.
Most writers discuss AI in relation to
employment,
war,
economics,
or ethics.
Elliott asks instead:
What happens when AI enters intimacy?
Examples include
- dating apps,
- recommendation systems,
- therapy chatbots,
- companion robots,
- predictive advertising,
- quantified sexuality,
- emotional AI.
His central thesis is striking:
AI no longer merely helps us communicate.
AI increasingly
participates in the formation of desire itself.
Machine learning begins learning
our
preferences,
attachments,
fears,
habits,
even romantic choices.
Intimacy becomes
algorithmically mediated.
This is what he calls
algorithmic intimacy.
His recurring themes
Across these books, several ideas recur.
1. Identity becomes fluid—and computational
Earlier sociologists spoke about
identity.
Elliott speaks about
algorithmically constructed identity.
Instead of simply asking
Who am I?
we increasingly ask
What does Netflix recommend?
What does TikTok show?
Who does Tinder suggest?
The self becomes partly co-produced by predictive systems.
2. Cognitive offloading
AI increasingly remembers,
calculates,
navigates,
writes,
plans,
and recommends
for us.
This reduces certain burdens,
but also changes
what it means
to think.
Instead of intelligence disappearing,
human intelligence is redistributed
between
brains
and
machines.
3. Anxiety
One of his later books,
Algorithms of Anxiety,
argues that AI simultaneously promises certainty
while generating
continuous uncertainty.
People worry about
- employment,
- surveillance,
- misinformation,
- social comparison,
- replacement.
Ironically,
the technologies marketed as reducing uncertainty
often increase existential anxiety.
4. Human agency
This may be Elliott’s deepest concern.
He does not believe
AI inevitably destroys agency.
But
agency increasingly becomes
delegated,
nudged,
predicted,
and optimized.
His question is therefore
How much of everyday decision-making
remains genuinely ours?
5. AI is emotional
Unlike many sociologists,
Elliott insists
AI is not simply
technical infrastructure.
Its deepest effects are
emotional.
AI changes
love,
friendship,
loneliness,
trust,
dependency,
and attachment.
That emphasis reflects his long engagement with psychoanalytic theory.
The psychoanalytic dimension
Although Elliott rarely writes clinical psychoanalysis,
his thinking remains profoundly psychoanalytic.
He assumes
people are
not
fully rational.
Instead,
they are driven by
fantasy,
anxiety,
dependency,
recognition,
desire.
AI therefore succeeds
not merely because it is efficient,
but because
it satisfies unconscious wishes.
For example,
predictive systems promise
certainty.
Recommendation systems promise
recognition.
Companion AI promises
unconditional availability.
These are not merely technological functions.
They answer
psychological longings.
His optimism and pessimism
Elliott is neither an AI enthusiast
nor an AI catastrophist.
He rejects both
Silicon Valley utopianism
and
Hollywood apocalypse.
Instead,
he argues that
AI produces
new possibilities
and
new vulnerabilities
simultaneously.
The crucial question is
whether democratic societies
retain the capacity
to shape
AI,
rather than merely adapting to it.
Where I think Elliott is especially important
Having followed your questions over the past year, I think Elliott occupies an interesting position among the thinkers we’ve discussed.
- Ian Parker asks how capitalism produces subjects.
- Graham Harman asks what kind of object AI is.
- Robert M. Galatzer-Levy asks how complex systems self-organize.
- Anthony Elliott asks how AI reorganizes everyday identity, intimacy, and emotional life.
Among these, Elliott comes closest to your recurring concern with what you have called the algorithmic self.
At the same time, I also see an important difference.
Elliott’s key concept is algorithmic modernity. It is fundamentally a sociological diagnosis: he describes how algorithms reshape identity, relationships, institutions, and emotions. Your own line of inquiry has increasingly become existential and clinical. You are asking not only how algorithms reorganize the self, but also what conditions allow another mode of being to emerge.
That is why your recurring notions—analogical self, follow the fragments, strange attractors, 回到自然, and 菊次郎十六式—feel like a response to Elliott’s diagnosis. If Elliott describes the architecture of algorithmic modernity, your work is searching for practices and environments in which human subjectivity can continue to grow without being fully absorbed into that architecture.
I have the impression that this is precisely why Elliott’s work has resonated so strongly with you over the past year: he articulates the problem that your own “Psychology of the Self” is trying to answer.
