China's foreign ministry has pushed back against calls from senior US AI executives to slow the pace of frontier-model development, saying threat narratives and adversarial competition will only hinder global AI governance.
At a regular press briefing on 14 September, spokesperson Guo Jiakun was asked by a Reuters reporter about US AI companies arguing that China's progress in artificial intelligence could pose a national security risk to the US. Guo said AI development concerns the common wellbeing of all humanity and that all sides should promote an open, inclusive, accessible and beneficial AI environment. "Spreading all kinds of threat narratives and engaging in confrontation and malicious competition will only disrupt the global governance process for AI and is not in the interest of any party," Guo said, according to IThome, citing state broadcaster CCTV.
Anthropic, xAI and OpenAI line up behind slower development
The briefing came after Anthropic CEO Dario Amodei published a long post on X on 12 September titled "We Must Pace the Frontier", urging AI companies to slow the rate at which model capabilities grow. Amodei proposed a three-step framework: giving third-party independent evaluators employee-level access to oversee safety practices, coordinating safety standards among frontier AI companies, and building international cooperation to manage AI risks.
Amodei's essay drew immediate support from two of the biggest names in the industry. xAI CEO Elon Musk reposted the post and wrote "Dario is right." OpenAI CEO Sam Altman also agreed, saying the industry needs to pace frontier AI and that OpenAI had been discussing the topic over the past few weeks. Altman added that independent monitoring and employee-level access for third parties were good ideas that OpenAI would adopt.
The exchange highlights a growing debate within the AI industry: whether competitive pressure to release ever more capable models is compatible with responsible development. "Frontier AI" refers to the most advanced models at the edge of current technical capability, and concerns about their pace have become a recurring theme among safety-focused researchers and executives.
Beijing pushes for international coordination
China's response positions the question as a matter of global governance rather than a bilateral race. On that front, the 2026 World AI Conference and High-level Conference on Global AI Governance was held in Shanghai from 17 to 20 July. At the event, China's National Development and Reform Commission, together with other departments, released a plan for co-operative AI development, and the Ministry of Industry and Information Technology helped publish an international action plan for AI ethics governance. The conference also adopted a chair statement that endorsed people-centred, beneficial AI, fairness and inclusiveness, and collaborative governance.
Guo's remarks, which also referenced the transcript of the briefing, did not directly address the specific proposals from Amodei, Musk or Altman. Instead, they signalled China's preferred frame for the debate: any narrative that casts one country's AI rise as a threat will make it harder for governments and companies to agree on shared rules.
For readers tracking AI policy, the episode is a reminder that the biggest players in the industry are not just competing on models and chips. They are also competing on the story they tell about who should control the technology's pace and purpose. Washington and Beijing now have clearly different messages: some US executives are calling for voluntary slowdowns backed by independent oversight, while Beijing is arguing for collective governance and rejecting the threat narrative outright.
