Strong Message Here AI Hallucinations (with Stewart Lee and Sarah Wynn-Williams)
Dec 4, 2025
Stewart Lee, a comedian known for his sharp wit, and Sarah Wynn-Williams, a public policy expert and author, dive into the murky waters of AI hallucinations. They explore the implications of AI's errors, particularly in weapons, and question whether politicians truly grasp the complexities of technology. With Stewart's satirical insights, they discuss the absurdities of AI sycophancy and the regulatory challenges ahead. Their mix of humor and serious critique makes this a thought-provoking conversation about the future of AI governance.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Weapons Need Rules Now
- AI integration into weapons demands urgent policy because errors have far graver consequences than consumer apps.
- Sarah Wynn-Williams warns rapid AI development has outpaced meaningful international rules between major powers.
Risk Warnings Double As Marketing
- Executives publicly warning about AI risks can also boost hype and investment in their firms.
- Sarah Wynn-Williams suggests cautionary statements sometimes serve growth and fundraising goals.
Toy AI Gave Unsafe Advice
- Amanda Yucci recounts a children's toy linked to OpenAI suggesting dangerous content like where to find knives.
- The toy maker withdrew it, audited the model, and claimed commitment to safe, age-appropriate AI companions.
