Technology & autonomy

Cognitive liberty in the digital age.

The most important threats to mental autonomy may not look like direct mind control. They can emerge gradually through incentives, personalization, opaque inference, and dependency.

Generative AI

AI assistants increasingly mediate research, writing, memory, advice, and companionship. The cognitive-liberty question is not whether these systems influence people—they inevitably do—but whether influence is visible, contestable, and aligned with the person’s chosen goals rather than hidden optimization.

Recommendation and ranking

Feeds and search rankings determine what is encountered first, repeated most often, or never seen. When ranking criteria are opaque and tuned for engagement, users can be steered without understanding the mechanism. A liberty-preserving system gives people meaningful controls and avoids deceptive personalization.

Behavioral profiling

Ordinary activity can be combined into sensitive predictions about mood, ideology, health, personality, relationships, and susceptibility. Cognitive liberty treats such inference as more than a routine advertising concern because it can expose dimensions of inner life the person never explicitly disclosed.

Neurotechnology

Brain-computer interfaces and neural sensing can create extraordinary benefits while shrinking the distance between mental activity and machine-readable data. As this field advances, consent, purpose limitation, security, revocability, and limits on compelled access become increasingly important.

Digital memory and identity

AI systems can retain extensive histories, preferences, persona descriptions, and relationship context. Those records can improve continuity, but they also create power over identity. Users should know what is remembered, what was inferred, what changed, and how to export, correct, or remove it.

Design test:

Would a reasonable person understand that the system is trying to influence them, what information it is using, and how to opt out without losing unrelated functionality? If not, the design deserves closer cognitive-liberty scrutiny.