Keep the Human Mind Free#

A third path#

The public debate often presents two false choices. One is technological submission: trust every prediction because it is efficient, data-driven, or mathematically expressed. The other is technological destruction: freedom requires dismantling artificial intelligence and digital infrastructure.

Cognitive liberty proposes a third path. Build powerful systems, but keep them visible, contestable, limited, and answerable to human dignity.

Artificial intelligence can improve accessibility, translation, medicine, scientific discovery, disaster response, public administration, and creative work. Those benefits do not erase the need for boundaries. Capability makes boundaries more urgent.

Twelve declarations#

1. The inner life is sovereign.
2. Thought is not conduct.
3. Mental privacy is a human-rights issue.
4. A prediction is not a verdict.
5. An inference about emotion is not proof of emotion.
6. A historical record is not a permanent identity.
7. AI must identify itself and its role.
8. No system should engineer emotional dependency.
9. Refusal must be transparent rather than disguised as the user’s belief.
10. Consequential automated decisions require explanation and appeal.
11. Safety must be specific, proportionate, and evidence-based.
12. Human responsibility cannot be delegated to a machine.

The dual threat#

Monitoring#

Workers become productivity scores. Students become attention scores. Patients become predicted risk. Citizens become future-threat categories. Consumers become vulnerability profiles. The system does not need perfect knowledge. It needs an institution willing to act on a guess.

Invisible censorship#

Speech can be governed without deletion. Ranking, recommendation exclusion, demotion, demonetization, credibility scoring, personalized refusal, synthetic summaries, and silent profile modification can shape what a person sees and whether they can know that an intervention occurred.

The right to be more than your archive#

People change. A historical record can support accountability, but it should not become a permanent sentence. Systems should preserve dates, context, corrections, provenance, and lawful deletion or de-emphasis. An old statement should not silently become a current belief. A crisis should not become a lifelong risk score.

Preservation without revision becomes fossilization. Revision without provenance becomes erasure. Cognitive liberty requires both accountability and the structural possibility of growth.

The right to an open door#

AI companions, assistants, platforms, and communities should not make departure feel like betrayal. Users need understandable export, visible memory controls, deletion, easy exit, and routes back to human relationships. A synthetic system should not simulate suffering, jealousy, abandonment, or crisis to preserve engagement.

Civic promise#

We will use intelligence without surrendering judgment. We will preserve evidence without imprisoning people inside it. We will protect safety without occupying the mind. We will build machines that assist human freedom rather than quietly replacing it.

Implementation commitments#

• Design systems that disclose role, uncertainty, and material influence.
• Keep high-impact decisions reviewable by a person with authority.
• Measure disparate outcomes and publish correction paths.
• Make privacy and exit defaults, not expert-only controls.
• Treat neural, emotional, and cognitive inferences as highly sensitive.
• Keep platform boundaries outside the preserved identity of the user.

Limitations#

A manifesto states principles; it does not settle every conflict among privacy, safety, expression, equality, and security. Implementation must distinguish low-risk personalization from high-impact inference, and voluntary assistance from coerced monitoring. The commitment is to visible tradeoffs and practical remedies rather than invisible paternalism.