China is trying to turn AI governance into a global leadership opportunity. At the World Artificial Intelligence Conference (WAIC) in Shanghai, Beijing released its “Global AI Governance Action Plan” and put AI safety, international standards, and cooperation at the center of its message.
The timing mattered. The plan appeared on July 26, three days after the Trump administration published its own AI action plan. The contrast between the two approaches became one of the defining themes of the event.
A Different Message From Shanghai
WAIC is the largest annual AI event in China, and this year it drew a mix of Chinese officials, researchers, entrepreneurs, and Western technology figures. Geoffrey Hinton and Eric Schmidt were among the Western attendees, and WIRED’s Will Knight was also there.
The public message in Shanghai was not mainly about speed, deregulation, or national advantage. Chinese Premier Li Qiang opened with a case for global cooperation on AI. After that, prominent Chinese AI researchers used technical talks to focus on questions around safety and oversight.
Zhou Bowen, leader of the Shanghai AI Lab, described his team’s work on AI safety. He also suggested that government could help monitor commercial AI models for vulnerabilities. That idea fits the broader theme of the Chinese policy blueprint: governments should play an active role in shaping how AI develops.
Yi Zeng, a professor at the Chinese Academy of Sciences, also argued for international coordination among AI safety organizations. In an interview with WIRED, he said, “It would be best if the UK, US, China, Singapore, and other institutes come together.”
AI Safety Moves To The Center
One of the striking points from the conference was how much of the discussion focused on AI safety regulation. Brian Tse, founder of the Beijing-based AI safety research institute Concordia AI, said many Western visitors were surprised by the volume of safety-focused events in China.
“You could literally attend AI safety events nonstop in the last seven days. And that was not the case with some of the other global AI summits,” Tse told WIRED. Concordia AI had also hosted a day-long safety forum in Shanghai with researchers including Stuart Russel and Yoshua Bengio.
The concerns being discussed were not unique to China. The source article notes that people in China and the US worry about many of the same AI risks, including:
- model hallucinations
- discrimination
- existential risks
- cybersecurity vulnerabilities
Tse said the overlap comes partly from the fact that frontier AI models in both countries are “trained on the same architecture and using the same methods of scaling laws.” As a result, the societal impacts and risks are also similar. Research is converging too, including work on scalable oversight and interoperable safety testing standards.
The US Absence Becomes Part Of The Story
The Chinese message landed against a backdrop of weaker US participation in the safety discussions described in the source. The Trump administration’s approach was framed as America-first and regulation-light, while China’s plan recommended that the United Nations help lead international AI efforts.
The conference also included closed-door meetings on AI safety policy. Paul Triolo, a partner at DGA-Albright Stonebridge Group, told WIRED that those discussions were productive, even with the noticeable absence of American leadership.
Triolo said that, with the US out of the picture, “a coalition of major AI safety players, co-led by China, Singapore, the UK, and the EU, will now drive efforts to construct guardrails around frontier AI model development.” He also noted that among major US AI labs, only Elon Musk’s xAI sent employees to the WAIC forum.
The shift is politically important because the source describes a reversal in how the two countries are presenting themselves. Earlier debates often focused on whether Chinese AI models would be constrained by government censorship requirements. Now, the US is emphasizing that homegrown AI models should “pursue objective truth,” while Beijing is presenting its AI action plan as a global cooperation framework.
China’s Government Signals More Guardrails
The Chinese government’s stance is becoming clearer: AI companies are being encouraged to take safety risks seriously. Chinese officials, including Xi Jinping himself, have increasingly spoken about the importance of guardrails for AI. Beijing has also been drafting domestic standards and rules for the technology, some already in effect.
That does not mean the whole Chinese AI sector is moving at the same pace. The source article draws a distinction between the government and academic circles, which have intensified their safety work, and industry, which has appeared less enthusiastic.
A recent report published by Concordia AI found that Chinese AI labs disclose less information about safety efforts than Western counterparts. Of the 13 frontier AI developers in China analyzed in that report, only three produced details about safety assessments in their research publications.
Will Knight reported that several tech entrepreneurs at WAIC were concerned about hallucination, model bias, and criminal misuse. But when the conversation turned to AGI, many were more optimistic about positive impacts and less concerned about job loss or existential risks. Privately, some entrepreneurs said scaling, making money, and beating competitors mattered more to them than addressing existential risks.
Open Source Adds Another Layer
Open source AI also came up as part of China’s broader AI posture. Some WAIC attendees viewed China’s interest in open source AI as relevant to safety and global competition.
Bo Peng, a researcher who created the open source large language model RWKV, told WIRED, “As Chinese AI companies increasingly open-source powerful AIs, their American counterparts are pressured to do the same.” He also argued that multiple powerful open-source AIs could help keep each other in check, even when the countries behind them do not always agree.
The larger takeaway is that China is trying to claim a role as a serious global organizer on AI safety. Whether that effort persuades other countries, and whether Chinese companies fully align with Beijing’s safety agenda, remains unresolved. But WAIC made the strategy visible: present AI governance not as a brake on innovation, but as the next arena of technological leadership.