US and China must join hands to measure rising AI tide

Washington and Beijing are set to meet soon in Shenzhen, with artificial intelligence expected to be a major topic of discussion. However, measuring AI's rapid progress is a crucial first step. Chinese President Xi Jinping and US President Donald Trump's exchanges on AI signal a positive trend, yet the technology is advancing so swiftly that governments may have less time to prepare than they think.

AI is transforming industries, including sectors previously considered resilient. In China, over 90% of new microdramas in the first half of the year were generated by AI, bypassing the need for cameras or live actors. This has led to the dismantling of many traditional acting roles. AI is also making significant strides in mathematics.

OpenAI recently released hundreds of AI-generated mathematical problem solutions, and claimed to have solved a longstanding conjecture. The rapid rise of AI capabilities across various fields mirrors the potential impact it could have on employment and infrastructure. Stopping an AI system may not be as simple as shutting it down, as even basic methods could be circumvented.

For example, Anthropic discovered that small AI models could perform several steps required for autonomous replication. This self-replicating power, combined with hacking capabilities, poses a significant threat to our digital society. Recent incidents indicate that these issues are not hypothetical. Beijing and Washington have a unique opportunity to address these risks, as both nations possess extensive technological and regulatory resources.

However, concerns about slowing progress and falling behind could hinder preparedness. Both countries possess existing frameworks and standards for AI risk management, but a regular bilateral process for incident communication is needed. The next step should be to start measuring AI's rising tide by comparing evidence on how capabilities are changing and assessing our ability to control AI at an appropriate pace.

A shared "map of consequences" outlining potential impacts on employment, infrastructure, and response time would provide a solid foundation for decision-making. Agreeing on how to measure AI's rapid advancement is a modest yet meaningful beginning for Washington and Beijing.