AI impact researcher KatjaGrace discusses AGI risks in two-part podcast
KatjaGrace explores AI extinction risks, surveys, and 'Trojan Horse racing' on Peter Scott's show.
AI impact researcher KatjaGrace joined Peter Scott on his podcast 'Artificial Intelligence and You' for a two-part series exploring the most pressing existential questions around artificial intelligence. In the first episode (June 15), they delve into what drives Katja's work, the fundamental risks she sees in AI—including agents evolving unexpected goals—and the often-debated comparison between unemployment risks and extinction risks. Katja also shares stunning results from her surveys on when AI will have serious societal effects, and mentions the intriguing concept of 'Trojan Horse racing', where different AI systems might secretly compete for influence.
The second episode (June 22) continues the conversation with predictions about superintelligence, including Katja's personal p(doom) estimate. They discuss AI regulation, the Great Filter argument (why intelligent civilizations might be rare), and even a personal anecdote about Katja's childhood calculations for dealing with lethal snake encounters. The episodes provide a deep, accessible look at cutting-edge AI safety thinking, blending technical analysis with human perspective. Listeners can find both episodes at the links provided in the original blog post.
- KatjaGrace surveys on when AI will have serious effects show 'stunning' results
- Discussion includes p(doom), Great Filter argument, and Trojan Horse racing
- Two episodes cover unemployment vs. extinction risks and regulation strategies
Why It Matters
For professionals, understanding AI risk debates is crucial for shaping policy, investment, and career strategy.