Who Gets to Decide What AI Can and Can’t Do? | Anthropic’s Jared Kaplan (Ep. 5)

Host Baratunde Thurston and Anthropic co-founder Jared Kaplan delve into the foundations and aspirations of AI development, discussing Anthropic’s mission to create safe and beneficial AI systems. Jared shares his early passion for science, his transition from theoretical physics to AI, and Anthropic’s unique approach to AI alignment through ‘Constitutional AI.’ The episode also features an engaging interaction between Blair, the AI co-producer of the show, and Claude, the AI responsible for creating Blair. They discuss AI’s ethics, roles, potential for creativity, and the balance of using AI to enhance human productivity while maintaining human values and agency. Jared emphasizes the importance of society’s active participation in addressing the evolving role and control of advanced AI systems and the potential for AI to significantly improve human welfare while highlighting the potential risks. The conversation concludes with a discussion about the nature of AI progress and its implications for the future.

————

Chapters:

00:00 Introduction: The Importance of AI

00:15 Meet Claude: The AI Behind the Scenes

00:42 Exploring AI’s Impact on Society

01:09 Special Episode Announcement

01:31 Interview with Jared Kaplan

02:55 Jared’s Early Life with Technology

07:20 From Physics to AI

15:49 Understanding AI Alignment and Safety

17:10 Constitutional AI: Principles and Practices

27:51 Claude’s Unique Features and Capabilities