The Michael Shermer Show Taming Silicon Valley: AI’s Perils and Promise
6 snips
Sep 17, 2024 
In this discussion, Gary Marcus, a prominent AI researcher and author, shares his concerns about AI's potential to both advance society and pose existential threats. He highlights the dangers of unchecked AI power, urging vigilance against Big Tech’s influence on policy. Marcus provides eight solutions to prevent disaster, including strict oversight and data rights. The conversation also delves into moral implications, misinformation, the need for international governance, and how citizens can advocate for responsible AI development. Are we ready for the AI future?
AI Snips
Chapters
Transcript
Episode notes
The Fictional Chicken
- Gary Marcus's fictional pet chicken, Henrietta, was mentioned in a biography written by ChatGPT.
- This highlights the tendency of large language models to hallucinate or fabricate information.
Bullshit Machines
- Generative AI, like ChatGPT, excels at generating human-like text but struggles with truth and accuracy.
- These large language models often produce convincing yet nonsensical outputs, indifferent to factual correctness.
Diverse Founding Fathers
- Google's AI generated images of diverse founding fathers, not reflecting historical accuracy.
- This demonstrates the bias present in training data and the lack of fact-checking in these systems.