US-China Cooperation Crucial for AI Safety Amid Growing Mistrust

Global AI safety initiatives hinge on cooperation between the U.S. and China, despite increasing skepticism between the two superpowers regarding each other's AI strategies. An upcoming meeting between the leaders of both nations in Washington is expected to address AI governance, a topic of growing urgency in both countries.

American tech leaders emphasize that averting the most severe AI risks necessitates some level of collaboration with China. However, the nations are actively competing for AI supremacy, with the U.S. president cautioning that regulations could benefit China. This dynamic creates a scenario where any agreements might be superficial, though potentially a starting point for larger dialogues.

Experts, such as Samm Sacks from Johns Hopkins School of Advanced International Studies, suggest that formal agreements may be nearly impossible. She argues, however, that leaders could create political space by acknowledging the risks AI poses to both nations. The U.S. has voiced concerns about China's ambitions in AI and has imposed restrictions on advanced AI chip exports to China.

Both countries are pursuing AI development through different avenues. U.S. companies are focusing on cutting-edge technology and data centers, while China aims for widespread adoption of open-source models. The U.S. is also considering developing its own open-source models to counter China's advancements, alongside accusations of technology theft.

Chinese officials have themselves expressed concerns about AI's potential threats, including cyberattacks and the undermining of political security. Both nations have observed that AI's emerging capabilities are raising alarms at the highest levels of government. While China has accused the U.S. of attempting to build an industry monopoly, U.S. representatives have called for dialogue and cooperation to manage safety risks.