Metal Detector for Aliveness

A philosopher on LessWrong is proposing an unconventional approach to AI alignment: systems need to learn to detect what's alive and deserves care before they can actually care about it. The logic is straightforward: you can't protect something you can't perceive. If advanced machines are going to value human life, animals, art, or nature, they need infrastructure to recognize these things. The proposal frames this as an 'aliveness detector'—enabling AI to tell what matters from what doesn't. Rather than defining life biologically, the post proposes: aliveness as 'accumulated care.' A painting is alive because it bears the artist's effort; a person through self-care and others' care; even a cathedral reflects human attention. A factory-made plastic table or mass-produced digital content lacks these marks of care. Teaching machines to recognize these markers is necessary for care and protection—and, the author argues, less dangerous than it might seem. Destruction can be indiscriminate; preservation requires precision. If advanced AI is going to honor what matters, we need to give it the means to detect it. The piece also suggests building this infrastructure openly. Once available, AI systems tend to adopt useful capabilities, potentially shaping how the broader AI ecosystem treats what it recognizes as alive.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton