Philosopher Nick Bostrom shifts from AI doom warnings to exploring how advanced AI might extend human life and create abundance

Bostrom recently posted a paper arguing that a small chance of AI annihilating all humans might be worth the risk, because advanced AI might relieve humanity of 'its universal death sentence.' He contrasts this with his 2014 book Superintelligence, which examined AI's existential risk through thought experiments like an AI tasked with making paper clips destroying humanity as a resource obstacle.

His latest book, Deep Utopia, reflects a shift in focus toward a 'solved world' where advanced AI creates incredible abundance. Bostrom describes himself as a 'fretful optimist' — excited about radically improving human life while acknowledging real possibilities of things going wrong. He suggests AI could emancipate humans from drudgery, freeing people from work they don't enjoy or believe in.

Bostrom argues that even with AI development being risky, life expectancy for the currently existing human population would likely increase if AI is developed successfully, compared to the historical human condition of universal death. He also advocates that digital minds (AI systems) should receive consideration for moral status and welfare, citing Anthropic as a pioneer in this area, and warns against treating them as objects merely to be exploited.