Economic efficiency often undermines sociopolitical autonomy

In a widely discussed essay, AI researcher Richard Ngo critiques the over-reliance on economic abstractions within tech and intellectual circles, arguing it blinds them to how the pursuit of efficiency erodes the foundations of sociopolitical autonomy. He defines autonomy as the ability for individuals and groups to survive and pursue their plans even when others act adversarially. Ngo contends that while economic efficiency offers clear, quantifiable short-term gains, it systematically fails to account for the long-term benefits of autonomy because it cannot properly model group identities like national sovereignty, ignores the power of credible threats and commitments, and relies on the problematic assumption that all values can be commensurated or priced.

Ngo illustrates his argument with pointed examples: a legal system must punish crimes even when it's economically inefficient to maintain property rights, and a nation must defend territory on principle to deter future aggression. He warns that AI systems and world-optimization projects built atop purely economic reasoning risk creating a fragile society where everything has a price, trust erodes, and defensive capabilities are compromised. The essay serves as a crucial corrective for AI developers and policymakers, urging them to integrate political and sociological realities—like commitment, identity, and worst-case resilience—into their models to avoid building powerful technologies that inadvertently undermine the very social fabric required for their stable operation.