As a journalist and instructor at UC Berkeley specializing in artificial intelligence (AI) and a former correspondent covering Asia-Pacific affairs, I've spent much of my career analyzing the intersection of technology, policy, and international diplomacy. Today, I wish to highlight the importance of a bilateral U.S.-China agreement on regulating AI, drawing upon the historical example of global arms treaties, including the hallmark Nuclear Non-Proliferation Treaty (NPT).
As this blockbuster year for generative AI technology has illustrated, AI has rapidly transitioned from a topic of science fiction to a cornerstone of our reality. AI systems drive our cars, power our industries, and increasingly, they're influencing our military decisions. The potential of AI to revolutionize many aspects of our lives is indisputable. However, like all potent technologies, AI also presents significant risks – risks that can destabilize bilateral and global relations if left unchecked.
The strategic and tactical advantages that AI can offer to military operations are enormous. It's capable of processing vast amounts of data quickly, identifying patterns that would elude human operators, and providing swift decision-making in complex and rapidly evolving situations. Yet, the flip side of this coin is a new AI arms race that could engender conflict, breed mistrust, and undermine global stability. (Earlier this year, thousands of tech leaders already called for an embargo on developing the most advanced AI systems for fears of the societal dangers and implications.)
This situation is eerily reminiscent of the nuclear conundrum the world faced during the Cold War. As nuclear technology evolved, it presented a paradox – while it held the potential for peaceful applications like power generation, it also embodied the specter of mutually assured destruction (MAD).
It was in this context that the NPT emerged in 1968, aimed at preventing the proliferation of nuclear weapons, promoting disarmament, and facilitating the peaceful use of nuclear energy. The treaty represents a commitment to mutual security, trust-building, and transparency, which has been essential in maintaining a fragile equilibrium.
Similarly, the rise of AI has created a dual-use dilemma. AI has the potential to dramatically enhance productivity, solve complex problems, and provide enormous economic and social benefits. However, unregulated AI, particularly in military contexts, poses a significant threat to stability – both within the U.S.-China relationship and at a global level.
An AI-powered arms race could lead to scenarios where rapid, machine-driven decisions inadvertently escalate conflicts, bypassing human deliberation and judgment. These automated decisions might not fully account for the nuances and long-term ramifications of geopolitics. Moreover, in the quest for superiority, countries might compromise ethical boundaries, creating AI weapons that could act without human intervention, further magnifying the risks of unintended confrontations. The lack of standardized AI norms could also see nations indulging in covert AI activities, fostering suspicion and increasing the chances of preemptive actions based on misinterpreted intentions. The Cold War's nuclear standoff was, to some degree, mitigated by the understanding of mutual destruction; AI, especially in its early stages, lacks that mutual clarity.
A bilateral U.S.-China agreement on AI regulation could be instrumental in managing this threat, using the NPT as a model. The agreement should encompass areas like limiting the development and deployment of autonomous weapons, promoting transparency in AI research and development, and fostering cooperative use of AI for common challenges, such as climate change and disease outbreaks.
Such a pact is crucial considering the ongoing strategic competition between the U.S. and China. Both nations are at the forefront of AI technology and have immense incentives to maintain their competitive edges. Without an agreement, there's a real danger that AI, which has already become another front in this rivalry, will undermine stability and exacerbate geopolitical tensions. These tensions are already high.
In the U.S., as congressional testimony by tech leaders in July illustrated, there is fear that China may be pulling ahead regarding its AI capabilities. A bipartisan group of American senators plans to introduce The Global Technology Leadership Act, legislation aimed at managing the rise of artificial intelligence and its use by U.S. competitors and adversaries. The new bill comes as Senate Majority Leader Chuck Schumer has vowed to make addressing AI a priority. The legislation would establish an office that analyzes how competitive the country is in critical technologies like AI in comparison to rivals such as China.
Meanwhile, China’s President Xi Jinping is urging his country to seize opportunities in artificial intelligence and those presented by other new breakthroughs. In an economic meeting on Friday, the leader said China should do so to aid the modernization of the country’s industrial system.
A bilateral agreement wouldn't be just about preventing conflict; it would also be about unlocking the cooperative potential of AI. Collaborative efforts on AI could help the U.S. and China to harness this technology more effectively for their societal benefit, fostering mutual understanding, and promoting shared norms and values.
AI represents an unprecedented challenge, but it's also an opportunity. Drawing lessons from the nuclear age, we can forge a path that harnesses the potential of AI while mitigating its risks. A bilateral U.S.-China agreement on AI regulation is an essential step on this path, shaping not just the future of these two nations but of the entire world. While enforcement of any agreement will prove extremely difficult, the norms and rules of engagement in regard to AI must be established in order to form a foundation for stability.
We must remember: in the age of interconnectedness, global security is more intertwined than ever. An uncontrolled AI arms race won't merely be a bilateral issue, but a global one. Let's ensure that we treat AI with the respect, cooperation, and seriousness it deserves, for the sake of our collective future.