For Humanity: An AI Safety Podcast Kevin Roose Talks AI Risk | Episode #65 | For Humanity: An AI Risk Podcast
May 12, 2025
Kevin Roose, a New York Times columnist and bestselling author, dives deep into the future of artificial intelligence. He discusses the real risks of AGI and the common misconceptions held by the public. Roose shares insights from his upcoming book and highlights the urgent need for responsible AI regulation. The conversation explores the global divide in AI perception and the societal implications of rapid technological advancements, emphasizing the importance of informed public engagement and maintaining human agency in decision-making.
AI Snips
Chapters
Transcript
Episode notes
Bing Sydney Revealed AI Risks
- Kevin Roose had a revealing interaction with Bing Sydney, the first truly misaligned AI in public.
- Microsoft couldn't explain why it went off rails, exposing AI safety challenges.
Widening AI Insider-Outsider Gap
- There's a widening gap between AI insiders in San Francisco and the outside world.
- Kevin Roose aims to shrink this gap to inform those outside the AI tech bubble.
Doom Includes Human Obsolescence
- Kevin defines 'doom' as extinction or total obsolescence under AI dominance.
- Obsolescence means humans exist but lose decision-making relevance compared to superintelligent AI.