U.S. Can Hold AI Talks with China Because ‘We Are in the Lead,’ Bessent Tells CNBC as Nations Plan Safety Protocol

The recent announcement of a protocol on best practices for artificial intelligence (AI) between the US and China has sparked a mix of reactions, ranging from cautious optimism to outright skepticism. At its core, this development rests on a dubious assumption: that the US is indeed “in the lead” when it comes to AI research and development.

US Treasury Secretary Scott Bessent’s remarks to CNBC suggest that the US’s supposed dominance in AI allows for these high-level discussions with China. However, this assertion belies a more complex reality. While the US has made significant strides in AI, particularly in large language models, other countries are quickly closing the gap.

China’s investments in AI have yielded impressive results, including advancements in natural language processing and improvements in computer vision. These developments have not gone unnoticed by the global community, and it’s increasingly clear that China is no longer a laggard in this field. Recent breakthroughs in Chinese AI research demonstrate that Beijing’s efforts are paying off.

Bessent himself mentioned the upcoming releases of Google’s Gemini and OpenAI’s large language models, which promise to bring about significant advancements in capabilities. This trend suggests that the competitive landscape in AI is far from settled. The notion that the US is “in the lead” may be premature, given the rapid pace of progress worldwide.

The real question is what this means for global security. The potential for non-state actors to exploit advanced AI technologies has long been a concern, and it’s unclear how effective these best practices protocols will be in preventing such misuse. As more countries invest in AI research, the risk of escalation and unintended consequences increases.

One area where progress is being made is in developing safety protocols for AI. The protocol announced by the US and China aims to address concerns around responsible AI development, including issues like data governance, transparency, and accountability. However, these efforts must be complemented by more concrete measures to prevent AI-enabled cyberattacks and other malicious uses.

Bessent’s comments also touched on Taiwan, a highly sensitive topic in Sino-US relations. While it’s true that Trump emphasized the importance of maintaining stability in this region, the implications of the US-China talks on AI remain unclear.

The assumption that the US is “in the lead” when it comes to AI may be a comforting fiction. As more countries join the AI race, policymakers must confront the very real risks and uncertainties associated with this rapidly evolving field. The announced protocol on best practices for AI represents a crucial step forward, but it’s only the beginning of a much larger conversation about how to ensure responsible AI development and mitigate its potential dangers.

The Treasury Department’s recent actions in Asia underscore the complexity of these issues. As policymakers move forward, they must prioritize transparency, collaboration, and a nuanced understanding of the AI landscape – rather than relying on simplistic assumptions about who is “leading” the pack.

Ultimately, the future of AI diplomacy will depend on our ability to navigate these challenges with sensitivity, awareness, and a willingness to learn from each other. The US and China’s efforts are a crucial step forward, but they must be accompanied by a deeper understanding of the risks and uncertainties associated with this rapidly evolving field.