The US Aims to Win the AI Race, But China Wants to Win Friends First

On July 23, the White House released “Winning the Race: America’s AI Action Plan.” Three days later, at a major AI conference held in Shanghai, China released its own “Action Plan on Global Governance of Artificial Intelligence” (人工智能全球治理行动计划). The timing may be coincidental, but the divergent paths the two leading global competitors on AI are taking are starkly evident from these two documents.

The US plan is 28 pages long, and is accompanied by three executive orders that implement some of the proposed actions in the document. The Chinese plan contains only 13 points, and could be fitted into 2 to 3 pages, and it was only published in the Chinese language.

These two plans could not be more different, and they clearly set out the different philosophies and priorities at play in the governments of the two countries about AI competition. The US approach, brashly put out in plain sight, is to “win the race” by making sure America will remain unassailably in first place. The plan calls for “exporting” the US AI technology stack in order to maintain dominance and secure the country’s economic and security goals. The titles of the three executive orders that accompanied the US AI action plan — “Promoting the Export of the American AI Technology Stack,” “Accelerating Federal Permitting of Data Center Infrastructure,” and “Preventing Woke AI in the Federal Government” — indicate that the US administration is more concerned with using AI to advance its ‘America First’ ideological agenda, promoting fossil fuel energy sources, and defending conservative viewpoints than it is about developing AI itself.

By comparison, China’s approach looks like a charm offensive. China pledges to share its technologies — its own Chinese AI stack — and its experiences in applying AI to all industries to create innovative ecosystems and data transfer norms together with other nations. China says it will work with other nations to establish governance processes, environmental rules, and safety guardrails. The plan is a display of China’s soft power at its best, written in high-minded language about building alliances for common destiny and prosperity, while the US has all but abandoned any such niceties. This is not to say that China is uninterested in competing for technology prowess and AI leadership, and in protecting its own national security. But its plan puts those concerns aside for another day. Here, China is focused on the geopolitical objective of luring the Global South countries, if not more, to adopt the Chinese model.

So, the purpose of this Chinese document is simply to put the best foot forward for external, international consumption. China’s own brutal internal strategy to win the AI competition can be found in others of its many planning documents and reports over the years, but this plan is all about global AI governance, for which the US administration appears to have little interest or patience.

Indeed, the word “governance” simply means “regulation” to many. The position of the current US administration is all about deregulation, or un-regulation, when there weren’t really any reglations to begin with. The US AI plan mandates that the federal government restrict AI-related federal funding to “states with burdensome AI regulations that waste these funds.” In fact, the word “governance” itself hardly appears in the US AI Action Plan, and when it does, in the section under Pillar III, “Lead in International AI Diplomacy and Security,” it is mainly just in reference to “countering Chinese influence in international governance bodies.” The rest of the section contains a hodgepodge of issues like exporting American AI, export controls on AI compute and semiconductors, national security risks in frontier models, and biosecurity. These are hardly about AI diplomacy. 

China wants to take over the void left by the Americans, by explicitly committing its support for a multilateral governance model for global AI cooperation, based on the United Nations’ Pact for the Future and its Global Digital Compact annex. It will support setting up future AI standards under the International Telecommunication Union, which, again, is part of the UN. For decades, China has been trying to pry away the governance and standard-setting process of the Internet from its present bottom-up, non-state-dominated multistakeholder governance model. It now sees a perfect opportunity to switch future AI standards and governance to be led by state governments, under the UN, with little to no say from the private sector and civil society.

If China gets its way, where can the Americans stand? The attitude of this US administration toward international governance bodies has been dubious and often inconsistent, and under Trump could be characterized as hostile. The US AI Action Plan complains of “Chinese influence” in these bodies, but apparently the US is more annoyed about “burdensome regulations” and “vague ‘codes of conduct’ that promote cultural agendas that do not align with American values.” Nonetheless, the plan recommends the US “leverage” its position “in international diplomatic and standard-setting bodies to vigorously advocate for international AI governance approaches that promote innovation, reflect American values, and counter authoritarian influence.”

This is one section of the US AI action plan that has received little attention from observers and the media, but actually represents the crux of US-China diplomatic and geopolitical competition on the future of AI governance. Unlike the Biden administration, the Trump White House has not expressed its support for the multistakeholder governance model, or even the UN and other multilateral bodies. If anything, it has not been supportive of the UN and its organizations, such as UNESCO, from which the US recently withdrew again.

But, the Trump administration apparently does see the difference in the importance of the ITU in the global AI race. After all, the various committees and working groups under the ITU have published over 150 AI-related standards so far, according to Doreen Bogdan-Martin, the ITU Secretary-General, who is an American diplomat, telecom specialist, and career ITU bureaucrat. In fact, Bogdan-Martin, elected to be the head of the ITU in 2022, was endorsed by Secretary of State Marco Rubio for a second term, a year ahead of schedule for the next election by ITU country state members.

Bogdan-Martin defeated her opponent, a former Russian telecom minister, in the ITU election in late 2022, several months after Russia’s invasion into Ukraine. That was hardly the best timing for a Russian candidate, even with Chinese backing. The fact that China continues to embrace the ITU as an institution does mean that, first, the country is confident of its ability to “play the game” at the ITU with sufficient support from like-minded member states, and, second, it may even feel reasonably confident in nominating a Global South candidate to run against the incumbent, thereby increasing its influence on the body. After all, with the Trump trade war on all nations and its attitude toward multilateralism, it is not inconceivable that the US will find itself in an uphill battle.