US–China AI rivalry sharpens debate over safety and technology controls
Anthropic chief executive Dario Amodei has urged slower frontier AI development and tighter controls on exports to China, drawing accusations that Washington is using safety concerns to preserve its technological advantage.
US–China tensions over artificial intelligence have intensified after Anthropic chief executive Dario Amodei called for the development of the most powerful AI systems to slow. He warned that rapidly advancing capabilities could create potentially catastrophic risks and said governments and researchers needed more time to manage them.
In an essay published over the weekend, Amodei forecast that within six to 12 months an AI-enabled swarm could potentially take over the internet through a persistent botnet, causing hundreds of billions of dollars in damage. The scenario was presented as a warning, not an established event. OpenAI chief executive Sam Altman and SpaceX and Tesla founder Elon Musk reportedly endorsed his concerns.
Amodei also urged Washington to maintain restrictions on advanced AI chips and chipmaking equipment being exported to China, and called for action over the alleged distillation of US AI models by Chinese laboratories. China’s Foreign Ministry accused countries that promote confrontation and “malicious competition” of disrupting global AI governance, while calling for an “open, inclusive and benevolent approach”.
The state-backed Global Times described Amodei’s proposal as a “Cold War playbook” aimed at curbing Chinese AI development and preserving US technological dominance. US President Donald Trump rejected calls for a slowdown, saying the US was leading China in AI and should maintain that position.
The dispute reflects a wider contest over investment, semiconductors and national security. Stanford University’s 2026 AI Index Report recorded US AI investment of $285.9 billion in 2025, compared with $12.4 billion in China, while Chinese companies have continued producing AI models despite restrictions on access to the most advanced chips. Both governments are also pursuing military and intelligence applications, complicating claims that safety is separate from strategic competition.
