Interview, Conference Presentation
China & AI: What we get wrong | Jeffrey Ding (2020)
- AI is projected to evolve from a potential unifying global challenge for major powers like China, the US, Russia, and others into a sector where rapid software scaling contrasts with slower physical infrastructure development, though central planning may prove counterproductive for complex technological progress.
- China's semiconductor capabilities are anticipated to eventually catch up to states existing five to ten years ago as technical iterations hit physical limits, potentially leading to a domestic industry that reduces foreign dependency and diminishing returns for the current 80,000 Hours-style "China specialist" narrative.
- The concept of AI governance may shift toward disaggregated subdomain metrics rather than broad indices like mobile user counts, with the US potentially needing to treat China as an "indispensable nation" while managing risks of scientific cooperation suspension or export control ambiguity.
- Social and economic shifts in China may include the abandonment of the 9-9-6 work culture due to workforce backlash, a transition from "techno-nationalism" assumptions to a reliance on global innovation networks, and the potential integration of the social credit system into high-tech AI-controlled mechanisms.
- Geopolitical strategies may involve China preferring a stable US government while also exploiting internal US conflict to weaken alliances, with potential US benefits from a stable US status quo, while China may prioritize stability in disinformation campaigns unlike Russia's preference for chaos.
- Future technological applications may focus on system-wide transformations in manufacturing, logistics, and the industrial internet, competing with Western firms like Siemens and GE, while facial recognition and data monitoring could expand beyond Xinjiang to specific neighborhood thresholds.
- Short-term study abroad programs in China may result in "enlightened nationalism" among students, while the "community of shared future" slogan may clash with the reality of relative gains, potentially evolving into a practical world government framework or remaining a rhetorical device.
- Risks include the "self-fulfilling prophecy" of a tech cold war driving internet splintering, the misuse of facial recognition for ethnic minority monitoring, and the potential for purchasing Russian AI bots to attack critics or manage disinformation, though the social credit system's evolution remains a vague claim in some contexts.
- Global cooperation on AI safety and ethics may increase, influenced by Western thinkers, potentially balancing accuracy, cost, and cybersecurity, while the distinction between private companies and government labs may blur or remain distinct depending on the research domain.
- The AI arms race narrative may be redefined to focus on specific measurable technological systems and applications rather than discrete weapon counts, with a potential shift toward cooperative standards, though the "indispensable nation" status remains contested by other rising global powers.