Possible How we regulate AI could decide the next century
30 snips
Nov 12, 2025 Hosts dive into the collision of AI with real-world regulations, highlighting new laws in states like Utah and California. They discuss how OpenAI self-limits advice in sensitive areas to manage liability. The conversation turns global, exploring China’s push for an international AI governance body and the U.S. risks involved. A fascinating study reveals that AI performs better with rude prompts, sparking debate on user behavior and training ethics. Ultimately, they question what kind of intelligence we truly want to foster.
AI Snips
Chapters
Transcript
Episode notes
Demand Clear AI Transparency
- Transparency about when systems are AI should be accurate and non-deceptive.
- Public and audited disclosures can help regulators understand AI usage and risks.
Use Targeted Licensing And Disclosure
- Use licensing selectively and require disclosure if a service is unlicensed.
- Define clear certification criteria for AI to determine when it must be licensed.
Compare AI To Average Human Performance
- Evaluate AI by comparing it to average human performance, not an impossible ideal.
- Deploying AI that outperforms the average practitioner can reduce overall human suffering.