China's Ministry of Foreign Affairs also rejected the slowdown argument, saying it "only impedes progress in global AI governance." Speaking at a regular briefing on Sept. 14, ministry spokesperson Guo Jiakun said, "Spreading threat narratives and engaging in malicious competition serves no one's interests."
Anthropic Chief Executive Dario Amodei had laid out his case for slowing AI development in a lengthy post on Sept. 12. Key industry figures including OpenAI's Sam Altman, Google DeepMind's Demis Hassabis and xAI's Elon Musk expressed agreement and support for slowing development and introducing government regulation. Microsoft joined the trend on Sept. 14, releasing a draft of new guidelines centered on the idea that "people matter more than AI." Microsoft plans to gather feedback and apply the guidelines to actual model development work starting next year.
Global Times, the English-language newspaper affiliated with Chinese Communist Party organ People's Daily, noted that Amodei mentioned China 12 times in his blog post, warning of the "risk of reducing AI security issues to a geopolitical zero-sum game rather than a matter for cooperation among major powers."
U.S. Democrats, however, say they will tighten AI regulation if they take control of the House in the midterm elections. According to The New York Times, former President Barack Obama said at a private Democratic fundraiser in Manhattan on Sept. 10 that AI "could become dangerous if not properly controlled," and proposed establishing a framework for discussing the issue.
Meanwhile, Chinese AI models now account for a majority of global usage, overtaking the United States. Japan's Nihon Keizai Shimbun, citing U.S.-based OpenRouter, reported that Chinese AI accounted for 55% of usage in June this year on a token basis, compared with 43% for U.S. AI. As of Sept. 7, the most-used large language model on OpenRouter across all AI models was "Hy4 Preview," an AI model from China's Tencent.
