Counterintuitive case: China's authoritarian brakes might make AI safer than US race
Expert Victor Shih says China's fear of power usurpation may lead to better AI control.
The conventional wisdom in the West holds that US-owned AI labs must beat Chinese labs in the race to AGI/ASI, and that if a winner must emerge, better it be US labs. The author questions this assumption, especially given the financial interests of those in power.
Victor Shih, a Chinese political system expert, provides a counterpoint: the CCP is terrified of AI being used to usurp its power. They see 'developing the brakes' as just as crucial as the AI itself. This means designating humans in every government agency and commercial entity to slow or stop rogue AI. The author concludes that if rogue ASI is a bigger risk than a corrigible one, China's emphasis on control might actually make it the safer leader.
- Western media defaults to fearing a Chinese AI win, but the author questions this bias.
- Victor Shih argues CCP's primary goal is retaining power, making them likely to build strong human oversight (brakes) into AI systems.
- If rogue ASI is more probable than a corrigible one, China's authoritarian approach might paradoxically be safer than the US's speed-focused race.
Why It Matters
Challenges the binary 'US good, China bad' narrative on AI safety, urging a nuanced look at governance.