Thomas Holland

Thomas Holland writes on artificial intelligence safety and governance for the Nakada Foundation to Save Humanity. His work covers the technical concepts underlying AI risk, from alignment and corrigibility to mesa-optimization and instrumental convergence, and translates them into terms accessible to policymakers, advocates, and the general public who need to understand what is at stake.

His writing addresses both the technical and the political dimensions of AI risk. The problem of artificial superintelligence is not a purely scientific one. It is a governance problem: a question of whether the international community can build the legal and institutional frameworks to ensure that systems exceeding human-level intelligence remain compatible with human survival and flourishing. Answering that question requires clear communication across disciplines.

The Foundation's editorial position is that AI safety communication has too often been directed inward, toward researchers who already understand the problem, rather than outward toward the political and cultural audiences whose engagement is necessary for governance to happen. Effective advocacy requires writing that is accurate, accessible, and frank about the scale of the risk without retreating into either false reassurance or paralyzing catastrophism.

Thomas Holland's work on this site represents the Foundation's commitment to building that kind of public understanding: grounded in the research, oriented toward policy, and written for readers who are willing to think seriously about what it means to govern the most consequential technology in human history.