AI Safety Should Focus on Loss of Freedom, Not Power Concentration

sethlazar · x · 2026-09-01

The author criticizes the implicit normative loading often attached to discussions of "concentration of power" in AI safety circles, tracing this view to STS crowds and Foucault. The author argues that concentration of power is not inherently bad (e.g., in liberal democracies, parenting, or corporate governance). What is actually bad is the unjustified use of power, procedural illegitimacy, or power wielded by those lacking proper authority. The author suggests that the real concern is the substantively unjustified diminishment of individual freedom by those without proper authority—a hallmark of illiberalism—and that advocacy should focus on individual liberty rather than lamenting power concentration itself.

Original post →

More from Safety

Safety channel →