Shenzhen’s AI ambitions meet the US–China safety divide

AI-generated illustration

Analysis

For Shenzhen, the debate over global artificial intelligence safety reaches directly into the businesses building the Greater Bay Area’s technology economy. Home to Tencent and Huawei, the city has a stake in both the development of AI models and the computing infrastructure needed to run them.

Yet hopes for a global approach to AI’s risks depend on cooperation between the United States and China – superpowers that seem only to be more skeptical of each other’s strategy.

AI governance is expected to be on the agenda at a meeting planned for next week in Washington between President Donald Trump and Chinese leader Xi Jinping, according to The Associated Press. The topic has gained urgency in both countries, with American tech leaders saying that averting worst-case scenarios for rogue AI depends on a degree of cooperation with China.

Both countries, however, are vying for a decisive advantage, with Trump warning that efforts to regulate the technology would help China. Any agreements might therefore be limited, although they could provide a start to something larger.

“A formal agreement or consensus may be near impossible, but they do not need to go that far,” Samm Sacks, a senior fellow at Johns Hopkins University, told AP. “Trump and Xi just need to create political space by acknowledging AI poses risks to both countries.”

Shenzhen illustrates why separating competition from cooperation will be difficult. Tencent’s Hunyuan models place it in the contest to develop and distribute AI, including through openly available models. Huawei’s Ascend computing technology puts it at the center of China’s effort to reduce dependence on American AI hardware.

For these companies, the international debate concerns access to chips, the distribution of models and the rules governing their use. Decisions framed as national security measures in Washington or Beijing can shape commercial opportunities in the GBA.

Competing paths to AI leadership

Trump has pushed back on calls to strengthen regulation, claiming that doing so could cause China to eclipse America. He has repeatedly declared that the nation that leads on the technology “wins.”

“We are leading China, and all others,” Trump said Monday on social media.

The two countries are moving the technology forward in different ways. U.S. companies are chasing AI frontiers with computing power supported by cutting-edge chips and rapidly expanding data center infrastructure. China is pursuing wide global adoption of its open-source, more affordable models.

Chinese models have made major advances and gained traction with American companies even as U.S.-led restrictions bar China from obtaining the world’s most advanced AI chips. While most American frontier models are closed-source, Treasury Secretary Scott Bessent said Tuesday that the U.S. needs more open-source models to counter China.

Washington has accused Beijing of “aggressive, malicious” efforts to extract capabilities from American AI systems. China has accused the U.S. of trying to suppress Chinese companies to build a “monopoly of the AI industry.”

Still, those with ties to the Chinese government see coordination as useful. At the Xiangshan Forum in Beijing, former Chinese ambassador to Washington Cui Tiankai said Tuesday that AI is an area where “we should go ahead and build up” dialogue.

“Because there is a need, a growing need, for cooperation,” he said. “So how can we manage the balance?”

Calls for cooperation and containment

American AI companies face a similar contradiction. Anthropic and OpenAI restrict access in China and argue that protecting the U.S. technological lead is a national security imperative. Both also advocate cooperation with China on safety.

In an essay last weekend calling for slower AI development, Anthropic CEO Dario Amodei said Washington should constrain China’s progress, including maintaining restrictions on advanced chip purchases.

But he added that “global pacing will require cooperation with China, the autocratic country with by far the most advanced AI capabilities.”

China’s Ministry of Foreign Affairs responded Monday that “fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance which serves no one’s interest.”

The state-run Global Times called his remarks “packed with containment provisions targeting China” and described them as a “‘Cold War playbook’ for the AI sector.”

Nvidia CEO Jensen Huang, whom Trump has consulted, has generally favored some chip exports to China as a way to create an American technology ecosystem and maintain a more open relationship, given the number of researchers there.

Meanwhile, increasingly powerful AI capabilities have alarmed officials in both countries.

China’s Minister of State Security Chen Yixin warned in an article Sunday of threats ranging from political and ideological security to cyberattacks. Singling out Anthropic’s and OpenAI’s models, he said they could be used to “weaponize” cybersecurity vulnerabilities and threaten critical information infrastructure.

That concern has a direct Shenzhen connection. Paul Triolo, a partner at DGA-Albright Stonebridge Group, told AP that a U.S. security firm’s recent discovery of AI vulnerabilities in WeChat was among the developments causing concern in China.

Tencent’s messaging and payments platform shows how questions about AI safety can extend into services used in daily life. For the GBA’s technology businesses, the debate encompasses the security of products already in use as well as the risks posed by future systems.

Shared risks, rival alliances

Chinese officials said after Trump’s Beijing summit with Xi in mid-May that the two countries had agreed to pursue dialogue on AI development and governance. Little progress has been made.

Neither country “can simply seal itself off from the other,” Lizzi C. Lee, a fellow at the Asia Society Policy Institute’s Center for China Analysis, told AP. Both governments have an interest in maintaining human control of AI.

“We already know that neither government is going to abandon its competitive AI strategy or pursuit of technological advantage.”

Both are also building alliances. In July, China launched the World Artificial Intelligence Cooperation Organization with 29 initial founding member countries, including Russia and Pakistan. This month, Xi urged BRICS members to accelerate a global AI governance framework and establish an open-source AI community.

Some observers see the organization as a response to the U.S.-led Pax Silica initiative, launched last year to strengthen AI supply-chain cooperation with countries including Japan, Britain and Australia.

Asked by CBS about expectations for the Trump-Xi meeting, Amodei said “the more long-term thing would be working together to put a speed limit on the rate of AI progress,” while acknowledging the difficulty.

A recent post by Chinese state media-linked account Yuyuantantian was similarly guarded: “Only after the United States first demonstrates that the safety rules are equally effective for its own model companies can substantive discussions between the U.S. and China take place.”

For Shenzhen and the wider GBA, that leaves an unresolved question: can the two powers agree on safety rules that each accepts as protection against shared risks, rather than a means of holding back a competitor?

* Reporting by Chan Ho-him (AP)