US-China Begin Building AI Hotline: What Are They Afraid Of?
The United States and China are racing to become the world’s leading power in artificial intelligence (AI). However, amidst this intensifying competition, Washington and Beijing are beginning to build a sort of AI “hotline” to address shared risks that concern both nations.
This “hotline” is not a red telephone in the literal sense. What is being designed is an official dialogue and a potential notification mechanism to be used when an AI incident escalates to the level of national security.
The idea emerged during discussions between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng in New York on Sunday (20/09/2026). The meeting took place ahead of a scheduled meeting between US President Donald Trump and Chinese President Xi Jinping in Washington on Thursday (24/09/2026).
Bloomberg reported that both parties agreed to establish an AI dialogue to seek a mutual understanding of the technology’s objectives and threats. Washington subsequently proposed a more specific idea: a notification mechanism between the two countries.
“The United States has proposed that both countries have a notification mechanism,” Bessent said, as quoted by Bloomberg. The forum will be known as the “U.S.-China AI dialogue.”
However, there is a significant distinction between these two developments. While the US and China have agreed to build an AI dialogue, the system for notifying AI incidents remains a Washington proposal, and it is currently unclear whether Beijing has agreed to it.
The question then arises: why do two nations engaged in such fierce competition to develop AI require such a channel?
The answer lies in the changing nature of AI threats. The competition is no longer just about who possesses the smartest models, the largest data centres, the most advanced chips, or the most valuable AI companies. The risks are moving into the realm of national security.
The Financial Times reported that Washington wants to discuss various risks, including uncontrolled AI agents, AI-driven cyber threats by non-state actors, and the potential use of AI in biological weapons. The US and China are expected to hold another meeting regarding AI in approximately two months.
“The most important thing right now is to start the conversation,” Bessent told the Financial Times. According to him, establishing this dialogue is the first step toward reducing the risks of rapidly evolving technology.
When AI Touches National Security
The New York Times reported that the discussed notification system is designed to function when an AI issue reaches a national security level. This makes the discussions far different from ordinary technology or trade dialogues.
Washington is concerned, among other things, that AI could amplify cyberattack capabilities, military prowess, and surveillance systems. Leading US AI companies have also published reports regarding their models being able to penetrate or bypass testing environments intended to restrict their behaviour.
In such a situation, the issue is not merely whether an AI is smarter than the previous generation. The concern is what happens when systems become increasingly capable of acting autonomously, exploiting digital vulnerabilities, or being used by certain parties to attack another nation’s infrastructure.
For this reason, Bessent emphasised the need for a shift from opacity towards transparency between Washington and Beijing.
“Moving from opacity towards greater transparency between the world’s number one and number two AI powers,” Bessent said, as quoted by The New York Times.
Transparency is vital because AI-based attacks can raise questions that are difficult to answer quickly. Who did it? Was it a state, a criminal group, a terrorist organisation, an independent actor, or even an automated system whose actions exceeded the intentions of its creators?
Errors in answering these questions could be dangerous when involving the US and China. An attack on power grids, financial systems, communications, or military installations could quickly be perceived as an act of an adversary, even if the true source remains unknown.
The Associated Press noted that both nations share overlapping concerns regarding cyberattacks, biological misuse, loss of control over AI, and threats to critical infrastructure such as financial, energy, and transport systems. However, the space for cooperation remains limited by deep strategic mistrust.
This is where the function of the AI “hotline” becomes apparent. It is not intended to stop the US and China from competing, but rather to reduce the possibility of an AI incident escalating into a much larger crisis due to misperception.
Source: jawawa.id (https://jawawa.id/newsitem/us-china-begin-building-ai-hotline-what-are-they-afraid-of-1789980976)
Originally published by Republika (https://news.republika.co.id/berita/tlpg7b451/aschina-mulai-bangun-hotline-ai-apa-yang-mereka-takutkan), translated from Indonesian