Intelligence (Artificial) in a Fallen World

According to a new essay on LessWrong by Joseph Babbo, medieval theology offers an unexpected framework for thinking about artificial intelligence risk. Babbo draws parallels between the theological concept of angels—beings whose choices are irrevocable—and advanced AI systems that can make decisions with far-reaching, irreversible consequences. The argument goes: just as a single angel's choice can distort reality forever, a sufficiently powerful AI system making a single fateful decision could ramify through the world without recall. Babbo suggests we're at an inflection point where present technological choices can shape the future in ways we cannot undo, making our epistemology—our fundamental beliefs about what's knowable and what's worth knowing—crucial to safe development.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton