Is the A.I. Race Real?
Summary
Host Lizzie O'Leary discusses the US-China AI relationship with Scott Senger, co-director of the China AI Initiative at Carnegie Endowment. They analyze why the "AI race" framing is flawed, contrasting US closed-weight models with China's open-weight strategy driven by export controls. The conversation explores differing risk perceptions, noting that China prioritizes societal stability over existential threats, and examines the potential for a new diplomatic channel to manage AI safety amidst broader geopolitical tensions.
Topics discussed
Intro: AI journalism challenges and Slate+ sponsorship
Is the US-China AI race framing useful?
Show introduction and guest background
Comparing US and Chinese AI industry structures
Open-weight vs closed models and export controls
Differing perspectives on AI risk and control
Policy convergence: California and China regulations
Public attitudes and the 'AI Plus' strategy in China
China's diplomatic goals and status positioning
Building diplomatic channels and past failures
Insulating AI safety talks from broader tensions
The role of chip export controls in AI governance
Limitations of the Cold War analogy for AI
Advocating for step-by-step diplomacy over treaties
Key signals to watch for in upcoming summits
Credits and sign-off
Listen ad-free on Castria