The Illusion of the AI Race: Rivalry between the US and China
This week, the United States and China, global leaders in artificial intelligence, began dialogues to mitigate emerging security risks. But can meetings between Donald Trump and Xi Jinping really result in a reliable global governance framework for AI? It all depends on how much these fierce rivals are willing to cooperate.
Trump seems to want to convey the idea that discussions are progressing. Before the meeting with Xi, Treasury Secretary Scott Bessent announced that both sides have already discussed creating a new AI security notification mechanism. If China agrees to participate, it would open a channel for both countries to send alerts when their AI systems pose threats or behave unpredictably. However, the proposal comes at a time when China may not be ready to trust the US in discussions about AI security that do not aim to limit its capabilities.
Mutual Distrust and Modest Expectations
Distrust is palpable. Trump has increased export controls to block China from accessing technologies that drive AI, and Congress is considering bills that would cut the country's access to more chips and hardware. Additionally, the US Department of Justice recently accused six Chinese AI manufacturers of technology theft, an accusation that China vehemently contests.
Despite the tensions, China does not seem to have taken offense at Trump's rhetoric, which insists that the US must dominate China in AI. However, China has yet to publicly acknowledge the proposed alert system. Even if it did, the plan seems superficial, as it does not involve technical experts capable of assessing the urgency of the described risks.
Sam Bresnick, a researcher at Georgetown University's Center for Security and Emerging Technology, pointed out that the absence of technical experts is a red flag. Having technology experts on both sides would allow for high-level discussions on risk assessment, especially in discussions about military risks.
The Problem of the AI Race
The narrative of an "AI race" between the US and China is seen as deeply flawed by many experts. Alvin Wang Graylin, a global technology strategist, argues that this race is unlikely to benefit anyone. The US is focused on an illusory finish line in frontier AI, while the adoption of China's open models is rapidly growing, offering cheaper alternatives that better protect data privacy.
Graylin suggests that the real threat to AI security does not come from a race between nations, but from non-state actors who may use smaller specialized models for nefarious purposes, such as cyber or biological attacks. The assumption of a US-China conflict hinders the global cooperation necessary to protect against this real threat.
Paths to Cooperation
To overcome this rivalry, Graylin proposes that the US could alleviate global fears about AI by sharing high-level intelligence and opening bilateral and multilateral channels on AI security. This could include restoring more restrictive technology export policies, focusing only on genuine threats, and adopting reasonable security regulations.
The hope is that, eventually, the US will abandon the narrative of "winning an AI race." After all, there is no way to compete to achieve a safer world. True security in the age of artificial intelligence will come from cooperation, not from unchecked competition.
Comments (0)
Comments are moderated and if they violate our Terms and Conditions of use, the comment will be deleted. Persistence in violation will result in a ban of your account.