The Trajectory Yi Zeng - Exploring 'Virtue' and Goodness Through Posthuman Minds [AI Safety Connect, Episode 2]
13 snips
Apr 11, 2025 Yi Zeng, a prominent professor at the Chinese Academy of Sciences and AI safety advocate, dives deep into the intersection of AI, morality, and culture. He unpacks the challenge of instilling moral reasoning in AI, drawing insights from Chinese philosophy. Zeng explores the evolving role of AI as a potential partner or adversary in society, and contrasts American and Chinese views on governance and virtue. The conversation questions whether we can achieve harmony with AI or merely coexist, highlighting the need for adaptive values in our technological future.
AI Snips
Chapters
Books
Transcript
Episode notes
Moral Reasoning vs. Rule-Based AI
- Yi Zeng distinguishes between rule-based AI ethics and moral reasoning.
- He argues that true moral AI requires understanding, not just rule-following.
Limitations of Current LLMs
- Current large language models (LLMs) are perceived as potential threats due to their limitations.
- LLMs lack intrinsic motivation and understanding of why they should follow rules.
Importance of Sense of Self
- Moral reasoning necessitates a sense of self, enabling agents to distinguish themselves from others.
- Cognitive empathy, rooted in self-experience, is crucial for altruistic behavior and moral intuition.
