Preliminary discussions held in New York on Sunday between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng have centered on a proposed “notification mechanism” for artificial intelligence incidents. The initiative, effectively functioning as an AI hotline, aims to alert both nations when an AI-related event escalates to the level of a national security threat. These talks precede a scheduled summit between President Donald Trump and President Xi Jinping in Washington later this week.
Bessent stated that Washington is pushing for a broader dialogue on AI, with a specific emphasis on national security and shared threats. “We think that, just like with any cross-border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world is very important,” Bessent told reporters. Trump’s science and technology adviser, Michael Kratsios, reinforced this stance, telling Fox News that the primary objective is “avoiding shared risks.”
While the specifics of what would trigger such a notification remain undisclosed, experts have welcomed the move. Scott Singer, co-director of the China AI Initiative at the Carnegie Endowment for International Peace, described the proposal as an “obvious win for both sides.” He emphasized that the relationship requires guardrails to protect citizens, calling it a “productive first step. “
Beijing has offered a cautious response. The state-run Xinhua news agency confirmed that He and Bessent engaged in “frank, in-depth and constructive exchanges” regarding economic and trade matters, including AI, but did not confirm whether China had accepted the US notification proposal. Both officials departed without addressing the media.
This proposed channel would join existing crisis communication lines between the two nations. A direct secure telephone link was established in 1997 by Presidents Bill Clinton and Jiang Zemin, followed by a military hotline in 2008 designed to facilitate rapid communication between defense officials during dangerous incidents. However, the effectiveness of these channels has faced scrutiny; US officials have occasionally criticized the military hotline for delayed responses, and diplomatic friction persisted in 2023 when Beijing declined a request from then-US Defense Secretary Lloyd Austin to speak after the shootdown of a suspected Chinese surveillance balloon.
The urgency behind the US proposal aligns with growing concerns from the tech industry regarding the pace of AI development. Earlier this month, Anthropic CEO Dario Amodei warned that AI capabilities could advance rapidly enough to potentially seize control of the internet via persistent botnets within six to twelve months, causing hundreds of billions in damages. His warnings were supported by rivals Sam Altman of OpenAI and Elon Musk of SpaceX. Despite these appeals for caution, Trump has maintained that the US must preserve its competitive edge, asserting that “whoever wins AI, wins.”
Although the United States currently leads in private investment and semiconductor access for AI, China is closing the gap. Stanford University’s 2026 AI Index Report indicates that US companies invested $285.9 billion in AI in 2025, significantly outpacing China’s $12.4 billion. Nevertheless, Chinese firms have demonstrated resilience against US chip restrictions, highlighted by the January release of DeepSeek, which was developed at a fraction of the cost of Western competitors and caused a drop in US tech stock prices. China also holds advantages in rare earth mineral production and electricity generation, providing cheap power for data centers.
The upcoming summit, scheduled from Wednesday to Friday, will address a wide range of contentious issues, including unresolved tariffs, sanctions on Russian energy affecting Chinese and Indian buyers, and China’s role as a key ally of Iran amid the ongoing conflict. Additionally, longstanding disputes over rare-earth dominance and Taiwan remain on the agenda. Experts suggest that the AI notification mechanism could serve as a rare area of potential agreement. Singer noted that while the 2024 dialogue failed because AI risks were entangled with broader geopolitical issues, establishing a dedicated channel now could be a significant breakthrough.
The botnet warning from Anthropic is terrifying. If this hotline actually prevents a panic-driven escalation, it is a genuine win for everyone.
DeepSeek proved China can innovate cheaper under sanctions. I doubt they will fully comply with US notification demands anytime soon.
Isn’t it ironic that the same leaders screaming about winning AI are now asking for crash brakes? The stakes are too high for games.
A hotline is a good start, but we need to remember the military channel failed during the balloon incident. Words are cheap.