Modern Wisdom #650 - Geoffrey Miller - How Dangerous Is The Threat Of AI To Humanity?
15 snips
Jul 6, 2023 In this insightful conversation, Geoffrey Miller, a professor of evolutionary psychology at the University of New Mexico, dives deep into the existential threats posed by AI. He discusses the potential for AI to outpace human intelligence, raising fears of a 'god or demon' scenario with Artificial General Intelligence. Miller unpacks the urgent need for ethical governance and public discourse surrounding AI's rapid evolution. He also explores the implications of AI companionship and emphasizes understanding both existential and suffering risks for a safer future.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Risk Landscape
- AI systems are becoming smarter and faster than humans, posing a potential threat.
- This speed advantage creates incentives to give AI agency, increasing the danger.
Narrow AI Dangers
- Even narrow AIs, without general intelligence, can be dangerous if misused.
- They could design bioweapons or create destabilizing deepfakes.
Unforeseen AI Capabilities
- Rapid advances in hardware and deep learning have led to unexpected emergent properties in AI.
- ChatGPT's capabilities surprised researchers, highlighting the unpredictable pace of AI development.
