{"slug": "the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline", "title": "The US Pitched China an AI Red Phone Modeled on the Cold War Hotline", "summary": "US Treasury Secretary Scott Bessent proposed a US-China AI incident notification mechanism during eight hours of talks with Chinese Vice Premier He Lifeng at JPMorgan Chase's Manhattan headquarters on September 20, modeled on the Cold War Moscow-Washington hotline set up after the 1962 Cuban Missile Crisis. Bessent, who called the session \"very successful,\" said the two sides would define shared categories of dangerous AI, including uncontrollable autonomous agents, non-state hackers and bioweapons risks, and plan to meet again in China within two months; the proposal now goes to Donald Trump and Xi Jinping, who meet in Washington this week. US Trade Representative Jamieson Greer said export controls on AI chips and semiconductor manufacturing equipment remain outside the dialogue, and Xinhua's readout of the talks mentioned AI but omitted the notification mechanism entirely.", "body_md": "*The US wants a dedicated line to China for the moment an AI system goes rogue, built on the same logic as the phone Washington and Moscow set up after the Cuban Missile Crisis. Beijing's own summary of the talks barely mentioned it.*\n\nTreasury Secretary Scott Bessent spent eight hours in a room at JPMorgan Chase's Manhattan headquarters on September 20 with Chinese Vice Premier He Lifeng, and came out proposing something that has nothing to do with tariffs: a notification mechanism that would let Washington and Beijing warn each other the moment an AI incident starts to look like a national security threat. Bessent called the session very successful. The idea now heads to Donald Trump and Xi Jinping, who meet in Washington this week.\n\nHere's the plain version of what Bessent is asking for. Not a treaty. Not an inspection regime. Just a phone line you pick up when something's already gone wrong. We want to open a communications line, an incident line, so that we have constant communications, especially in the event of some kind of an incident, Bessent told reporters, according to Reuters. The two governments would also work out shared definitions of what actually counts as dangerous: uncontrollable autonomous agents, non-state hackers, bioweapons risks. Bessent has already named the arrangement the USA-China AI dialogue, and says the two sides plan to meet again in China within two months.\n\n## The Cold War playbook\n\nThe Cold War comparison isn't a flourish someone added for effect. It's the model. After the Cuban Missile Crisis in 1962 nearly tipped into nuclear war over a communications breakdown, the US and the Soviet Union built the Moscow-Washington hotline for exactly one reason: so a misread signal wouldn't spiral into something neither side wanted. According to Axios, officials on both sides are now drawing that comparison explicitly, treating a rogue autonomous AI system, or an AI-driven action that could be mistaken for an attack, as the modern version of the same problem. Two nuclear powers need a way to say that wasn't us before anyone reaches for a bigger response.\n\n## What's missing from the readout\n\nOne thing is deliberately not on the table. US Trade Representative Jamieson Greer made clear that export controls on AI chips and semiconductor manufacturing equipment sit outside this dialogue entirely, according to CNN. That's the single issue Beijing has pushed hardest to get reopened, and Washington is keeping it walled off in its own lane. This channel is about incidents and escalation, not about who gets to buy Nvidia chips.\n\n[Washington threatens to sanction Chinese AI firms over distillation theft as Beijing fires back](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-firms-over-distillation-theft-as-beijing-fires-back/)\n\nUS Treasury Secretary Scott Bessent threatened sanctions and Entity List blacklisting against Chinese AI firms over industrial-scale distillation campaigns, citing forensic evidence of American model watermarks inside Chinese products. China's commerce ministry called it 'AI hegemonism' Monday and threatened countermeasures, pointing out that US... - [how to use Chinese AI models](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-firms-over-distillation-theft-as-beijing-fires-back/) - [AI sanctions China trade war](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-firms-over-distillation-theft-as-beijing-fires-back/)\n\nThen there's the asymmetry that actually makes this story interesting. Xinhua, China's state news agency, published its own readout of the same eight hours. It mentioned AI. It never mentioned the notification mechanism at all, describing instead frank, in-depth and constructive exchanges on trade and investment. That's not necessarily a rejection. State media in Beijing rarely confirms a US framing before its own leadership signs off, and He Lifeng didn't shoot the idea down in public either. But it does mean the proposal Bessent is carrying into this week's summit is, for now, a one-sided announcement. Nobody on the Chinese side has said yes on the record.\n\nThat gap matters more than it looks. A crisis hotline only works if both parties actually pick up the phone. The US-Soviet line succeeded because both governments treated it as insurance against an accident neither wanted, not as a favor to the other side. If Beijing decides this is a mechanism the US built mainly for its own protection, it could nod along at the summit and give it no teeth back home. Trump and Xi meet in Washington this week. Whether Xi says yes to this on the record is the real test, not whatever got said in a conference room in New York beforehand.\n\nFor an industry that's spent three years arguing over export bans and chip smuggling routes, this is a different kind of AI story. It isn't about who owns the compute. It's about what happens when an autonomous system does something neither government intended, and whether the two biggest AI powers on earth have even agreed on who calls first.\n\n**Also read:** [Oklo's Reactor Hit Criticality, and Its Stock Is Riding an AI Power Rally](https://startupfortune.com/oklos-reactor-hit-criticality-and-its-stock-is-riding-an-ai-power-rally/) • [China Probes DeepSeek and Moonshot Over Secret Data Routing to Claude](https://startupfortune.com/china-probes-deepseek-and-moonshot-over-secret-data-routing-to-claude/) • [OpenAI Pitches a US-Led AI Rulebook to Counter the UN's Watchdog Plan](https://startupfortune.com/openai-pitches-a-us-led-ai-rulebook-to-counter-the-uns-watchdog-plan/)\n\n*This article is posted in [AI News](https://startupfortune.com/category/ai/), check it out for more related stories.*\n\n## Join the discussion\n\n[Open in the community →](https://startupfortune.com/community/)\n\nAlmost there. Sign in and your reply posts straight away.\n\n[Washington threatens to sanction Chinese AI models but American startups are already switching sides](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-models-but-american-startups-are-already-switching-sides/)\n\nTreasury Secretary Scott Bessent threatened sanctions against Chinese AI labs including DeepSeek and Moonshot AI over alleged IP theft via distillation attacks. But with US companies now running nearly half their AI workloads on Chinese models thanks to price gaps as wide as 36x, Washington's leverage is more complicated than the announcement... - [chinese ai models cheaper than american](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-models-but-american-startups-are-already-switching-sides/) - [american startups switching to chinese ai](https://startupfortune.com/washington-threatens-to-sanction-chinese-ai-models-but-american-startups-are-already-switching-sides/)", "url": "https://wpnews.pro/news/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline", "canonical_source": "https://startupfortune.com/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline/", "published_at": "2026-09-23 04:04:02+00:00", "updated_at": "2026-09-23 04:22:39.776769+00:00", "lang": "en", "topics": ["ai-policy", "ai-safety", "artificial-intelligence"], "entities": ["Scott Bessent", "He Lifeng", "Xi Jinping", "Donald Trump", "JPMorgan Chase", "Xinhua", "Jamieson Greer", "US Treasury"], "alternates": {"html": "https://wpnews.pro/news/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline", "markdown": "https://wpnews.pro/news/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline.md", "text": "https://wpnews.pro/news/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline.txt", "jsonld": "https://wpnews.pro/news/the-us-pitched-china-an-ai-red-phone-modeled-on-the-cold-war-hotline.jsonld"}}