The Behavioral Design Podcast · Samuel Salzer and Aline Holzwarth

Our Most Controversial AI Opinions – Season 4 Finale

·47 min·3 clips

The bigger risk is actually all the ways that AI goes right.

1. The Behavioral Design Podcast season 4 finale centers on AI opinions from Aileen Holsworth and Samuel Seltzer. 2. Aileen is a health tech advisor specializing in AI and product design, and Samuel is a behavioral strategist who founded Behavior Bites. 3. The episode asks what the hosts actually think about AI after a season of guest conversations and applied work. 4. Samuel opens with Nathan Fielder's The Rehearsal, which he links to exposure therapy and behavioral science. 5. He describes the season 2 focus on pilots and co-pilots, plus the claim that miscommunication contributes to plane crashes. 6. Aileen says Fielder breaks down power dynamics, incentives, and the behavior of pilots and first officers. 7. The hosts connect that work to AI by saying any human quality can be learned or emulated. 8. Samuel argues that the Turing test is no longer the right benchmark and that the episode is more about what AI can do now. 9. Aileen says vertical AI applications are already matching humans on many benchmarks. 10. They also say hallucinations have not really decreased in newer models and may be a feature, not a bug. 11. Aileen says the bigger risk is "all the ways that AI goes right," because helpful systems can create hidden backfire effects. 12. She warns that people may offload too much cognitive effort to ChatGPT and stop writing or thinking critically. 13. Samuel describes a personal Claude-based Oracle that gives him advice when he feels he has failed himself at night. 14. He says the system reflects his own values, including stoicism and Buddhism, but worries about scale. 15. The hosts discuss survey-style guest answers and say the season skewed more optimistic than pessimistic about AI. 16. They note a split on AI companionship, with Laura DeMolier and Peter Slattery positive and Kerry Morwetch and Sierra Greer skeptical. 17. Aileen says she is anti-AI companion and worries about teenagers forming strong bonds with AI. 18. The tone is conversational, reflective, and often blunt, with rapid back-and-forth and a few jokes. 19. Listeners interested in AI, behavior change, and product design. 20. Listeners looking for a technical deep dive into model architecture.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The Behavioral Design Podcast