Attention and agency

Persuasive technology and cognitive liberty.

Persuasion is part of ordinary communication. The cognitive-liberty problem begins when systems use hidden, individualized knowledge to exploit vulnerabilities and make refusal difficult.

Direct answer

Persuasive technology respects cognitive liberty when influence is visible, understandable, proportionate, and easy to refuse. It threatens cognitive liberty when optimization is covert, based on intimate profiling, designed to bypass reflective choice, or embedded in a relationship where the user cannot realistically leave.

Persuasion, nudging, and hypernudging

A general argument is open to reflection and rebuttal. A nudge changes the choice environment, such as setting a default. Hypernudging uses continuously updated behavioral data to personalize that environment for a specific person. The system can alter timing, framing, repetition, social proof, and emotional cues according to inferred vulnerabilities.

What makes influence coercive?

Factors that increase cognitive-liberty concern
Lower concernHigher concern
General message visible to everyonePersonalized message based on intimate hidden inference
Clear commercial or political sourceUndisclosed sponsor or synthetic persona
Easy opt-out and practical alternativesDependency, lock-in, or penalty for refusal
One-time suggestionContinuous adaptive optimization
Supports a stated user goalOptimizes an institutional goal against the user’s interest

Attention is part of cognitive agency

Infinite scroll, autoplay, intermittent rewards, outrage ranking, and personalized notifications can shape attention before a person makes a conscious decision. For adults, the goal should be control and transparency rather than paternalistic prohibition. For minors, stronger limits on manipulative amplification may be justified while preserving deliberate information seeking.

Design requirements

  1. Disclose when content, ranking, or timing is personalized to influence behavior.
  2. Do not use sensitive mental, health, political, religious, or vulnerability inferences for covert persuasion.
  3. Offer a non-personalized or chronological mode that is genuinely functional.
  4. Separate the user’s stated objective from the platform’s engagement or revenue objective.
  5. Make dark patterns, deceptive interfaces, and artificial urgency actionable consumer-protection issues.
  6. Audit for disproportionate effects on children, distressed users, and people in coercive institutional relationships.