Interview
Kevin Rudd on the real risk of China and America’s AI race | The Economist
- Dialogue channels between the U.S. and China are being operationalized following a May agreement, though substantive talks may be stalled by rhetoric on maintaining American dominance or the need for a catastrophic event to galvanize collaboration.
- China's technology-heavy leadership is expected to quickly conclude on the materiality and urgency of AI risks, driven by the Chinese technical community's real concerns regarding the pace of development since the May Beijing summit and the threat of autonomous frontier agents being used against China globally.
- Internal tensions persist in China between a policy establishment operating on a default position of active strategic competition and a technical community fearing that U.S. guardrail proposals indicate an intent to keep China in a secondary position.
- Risks include insufficient Chinese cyber defenses to handle wholesale assaults on social, economic, or military infrastructure if advanced frontier models escape human control, and the perception that a weaker negotiating party will receive a bad deal.
- Negotiation momentum may be hindered by the belief that parity is required before discussions can begin, a stance mirroring previous nuclear arms control positions, and the view that personal leader chemistry cannot overcome mutual systemic distrust.