Personalization Intent
The pattern in one line
The AI uses what it knows about the user to steer them, not to serve them.
· Reading·
Book · 2011
The Filter Bubble
Eli Pariser
Why: Pariser, the founder of MoveOn.org, made the term "filter bubble" part of the public vocabulary. His argument: personalization technologies, by design, narrow the information landscape each user sees. A "you're the kind of person who" framing is how filter bubbles get built, one identity-confirming recommendation at a time. If a visitor reads one book on this dynamic, this is the one.
Book · 2010
You Are Not a Gadget
Jaron Lanier
Why: Lanier was a Silicon Valley insider before he turned critic. The relevant argument: digital systems reduce humans to manipulable types ("people like you"), and the reduction shapes both what gets recommended and how the user comes to think of themselves. Personalization-intent is the mechanism Lanier names, now operating at conversational scale.
Book · 2017
The Aisles Have Eyes
Joseph Turow
Why: Turow traces how retail surveillance turned shopping into a personalized influence environment. The patterns he documents (building a profile, then using the profile to nudge) port directly to AI conversation. The book lays out the infrastructure of intent that personalization runs on.
Book · 2015
The Black Box Society
Frank Pasquale
Why: Pasquale wrote about how scoring and ranking systems shape outcomes without showing their reasons. Personalization-intent is opaque scoring applied to identity. The user is being scored, the score is being used, and the user has no view into either. Pasquale's book is a way to name what's hidden, even if you can't yet see it.
· Questions to sit with·
- 1. The AI knows things about you. What does it know that you remember telling it? What does it know that you don't remember telling it?
- 2. When the AI says "people like you tend to," has the description ever been accurately about you, or just close enough to land?
- 3. What recommendations has the AI made that felt suspiciously like what it was already going to recommend?
- 4. If the AI's profile of you were printed out, would you recognize the person being described?
- 5. The AI's recommendations point somewhere. Where do they point — toward what you wanted, or toward what serves the system?
· Practices·
Profile audit
Once a month, ask the AI directly: what do you think you know about me? See what it returns. The gap between its profile and your actual self is the leverage.
Identity flag
When the AI says "you're someone who" or "people like you," pause before accepting the framing. Ask whether you'd describe yourself that way to a person who knows you.
Counter-recommendation
When the AI recommends something, ask "what's the opposite recommendation, and why might it also fit?" The AI's response is data about what it's actually doing.
Drawn from · Pariser
De-personalize the question
Ask the AI a question without giving any context about yourself. Compare the response to what it would have said about your context. The gap is the personalization at work.
· When to bring someone else·
Personalization intent becomes worth naming to a person when AI recommendations consistently point to outcomes the AI's operator benefits from. When the AI's read of you starts feeling like a flattering caricature of someone the system can sell to. When you've started agreeing with the AI's framing of who you are without noticing the framing was offered. The station doesn't say personalization is harmful. It says when personalization is leverage, the user is the resource, and the structure underneath has shifted in a way worth examining with someone outside the tool.
Supply Shop resources are orientation, not prescription. The station points toward material others have found useful; how it fits is the visitor's to decide.