To answer your question precisely: Anthony Elliott does not use the exact phrase “algorithmic self” as a fixed technical term (like some digital theorists do), but his work—especially in Algorithmic Intimacy and Making Sense of AI—clearly develops a theory of the self under algorithmic conditions.
Let me reconstruct his position carefully.
1. From “self” → “algorithmic modernity”
Elliott argues that we now live in what he calls “algorithmic modernity”—a phase where:
- predictive algorithms organize institutions
- and penetrate everyday life and subjectivity
This already implies a shift:
the self is no longer primarily formed by social norms or narratives
→ but by data-driven systems and automated prediction
2. The core idea: algorithms move
inside
the self
A key move in Elliott’s thought is this:
- algorithms don’t just influence us externally
- they become internal to how we decide, desire, and relate
He explicitly says that:
- machine intelligence learns our preferences
- and “discreetly shapes our personal behaviour”
👉 This is crucial:
The self is no longer sovereign → it is co-produced with algorithmic systems
3. Algorithmic intimacy = the laboratory of the self
Elliott develops this most fully through the idea of algorithmic intimacy:
- dating apps
- social media
- therapy chatbots
- self-tracking devices
These technologies:
- guide who we love
- how we feel
- how we understand ourselves
In his formulation:
intimacy today increasingly involves “smartphones, social media, softbots, self-tracking technologies”
👉 So the self becomes:
- relational
- but mediated by code, prediction, and data flows
4. The emerging “algorithmic self” (implicit in Elliott)
From these ideas, we can extract what you would call the algorithmic self:
(1) Outsourced decision-making
- choices (dating, consumption, even therapy) are delegated to algorithms
- this creates what Elliott sees as a crisis of agency
(2) Predictive shaping of desire
- algorithms don’t just reflect preferences
- they pre-structure what we will want
👉 desire becomes:
- anticipatory
- pre-coded
- statistically guided
(3) Continuous self-monitoring
- quantified self, tracking, feedback loops
→ the self becomes a data object
(4) Hybrid human–machine subjectivity
Elliott suggests a gradual:
“hybridization of human beings with automated systems”
👉 The self is no longer purely human
→ it is partly machinic
5. The danger: loss of autonomy and depth
Elliott is not celebratory.
He warns that:
- people become entranced by algorithmic systems
- and increasingly align their lives with machine logic
This leads to:
- weakened autonomy
- standardized identities
- shallow forms of connection
👉 intimacy becomes:
- efficient
- optimized
- but flattened
6. A short formulation (in your language)
Let me translate Elliott into a compressed philosophical form you might appreciate:
The algorithmic self is
a self that no longer chooses,
but is chosen in advance.
Or more analytically:
- The modern self = narrative + reflexivity
- Elliott’s self = prediction + delegation + data feedback
7. A bridge to your thought
Now, if I may connect gently to your ongoing project:
Elliott’s “algorithmic self” sits very close to what you called:
「演算法的我 vs 類比的我」
In your terms:
- Elliott describes the algorithmic self
- your work seeks the possibility of an analogical self
And even further:
- Elliott → “being shaped by algorithms”
- you → “為岩中花樹所見” (being seen by things)
👉 One could say:
- Elliott diagnoses the capture of the self by algorithms
- you are searching for a release of the self into the world (物的歸位)
8. Final condensation
If I compress Elliott into one sentence:
The algorithmic self is a form of subjectivity in which personal identity, desire, and relationships are increasingly shaped, predicted, and governed by data-driven systems, leading to a partial outsourcing of agency to machines.
If you want, we can take one more step:
👉 Compare Anthony Elliott vs Bernard Stiegler vs Byung-Chul Han
on this same question of the algorithmic self.
That would be a very sharp triangle.
