Policy researcher Adam Thierer discusses his critique of effective altruism and the use of numerical estimates for speculative AI risksAI safety is the field and practice of reducing harmful failures, misuse, loss of control, and unintended consequences from AI systems.. He argues that safety requires continuing experimentation rather than a single centralized solution.
Thierer describes concerns about proposals to regulate computation and AI developmentGovernance is the system of responsibilities, decision rules and oversight used to direct an organization or technology. AI governance specifies who can approve uses, set boundaries and respond when systems cause problems.. The interviewer offers a distinction between technical alignment and defense in depth, and notes changes in Nick Bostrom’s stated views.
Watch on YouTube
