Essays
Safety & Governance
Least-restrictive safeguards, due process, function creep, child protection, reversible authority, privacy-preserving safety, and evidence that interventions actually reduce harm.
How democracies drift toward censorship
Legitimate harm prevention can become information control when scope and review fail.
Personalized censorship is harder to see
Ranking, identity, profiling, and AI mediation can create user-specific information boundaries.
Identity before expression
Mandatory identity changes the cost of reading and speaking, while eligibility does not always require a full identity dossier.
Curiosity is not intent
Why studying dangerous ideas should remain distinct from requesting concrete assistance to cause harm.
Safety without surveillance
Privacy-preserving architecture can reduce harm without making permanent records of everyone’s lawful inquiry.
The safety poison pill, narrowed
What survives after adversarially testing the claim that safety rationales become a poison pill for cognitive liberty.
The expansion ratchet
Why exceptional powers can expand beyond their original purpose—and why that risk is serious without being inevitable.
Cognitive liberty for minors
Protect young people from exploitation and manipulative amplification without eliminating confidential inquiry.
Audit the gatekeeper
A ten-category method for testing whether AI and information systems actually respect cognitive liberty.
Against Nanny AI
Transparent boundaries and adult agency instead of covert paternalism.
Explore by question
Follow a topic, not just a chronology.
The topic map groups essays by recurring questions and shows where evidence, ethics, power, institutions, and cognitive liberty overlap.