The growing strategic importance of artificial intelligence is pushing the United States and China toward a form of communication that neither side has yet managed to establish on a lasting basis. During economic talks in New York, US Treasury Secretary Scott Bessent proposed creating a bilateral mechanism through which Washington and Beijing could notify each other about artificial intelligence incidents serious enough to create national security concerns. The proposal is significant because it seeks to create communication around risks without requiring either country to abandon its competition for technological leadership.
The proposal comes at a time when artificial intelligence has moved beyond being primarily a commercial and technological issue. Advanced systems are increasingly relevant to cybersecurity, military capabilities, intelligence, critical infrastructure and economic competitiveness. That expansion creates a practical problem for both governments: an incident involving artificial intelligence could cross national borders or be interpreted as hostile action before officials have established what actually happened. A direct notification channel could provide a way to reduce that uncertainty.
Competition Is Creating a Need for Communication
Washington and Beijing remain strategic competitors in artificial intelligence, particularly in advanced computing, semiconductor technology and increasingly capable AI systems. The United States has maintained restrictions on exports of advanced AI chips and semiconductor manufacturing equipment to China, while Chinese companies have continued developing domestic alternatives and increasingly capable models. The two countries therefore have strong incentives to protect their technological advantages while limiting risks associated with uncontrolled competition.
That combination makes conventional cooperation difficult. Any discussion about artificial intelligence can quickly overlap with national security and technology restrictions. The latest proposal attempts to separate one narrow area of cooperation from the larger technology rivalry. Instead of seeking agreement on how quickly AI should develop or which country should lead, the proposed mechanism would focus on incidents that could create serious security consequences.
This distinction matters because the two governments do not need to agree on the future of artificial intelligence to recognise that certain incidents could become dangerous through misunderstanding. A major cyberattack involving an AI system, for example, could be interpreted differently by the country experiencing it and the country believed to be responsible. Communication could potentially help establish facts before political or military responses escalate.
The Proposed System Addresses an Information Gap
Bessent described the idea as a notification mechanism covering AI-related incidents that rise to the level of national security. The broader proposal also includes an AI dialogue intended to establish greater transparency between the two leading AI powers. The precise structure, thresholds and obligations of such a mechanism have not yet been publicly defined, and China's public response has so far been limited.
That lack of detail is important. An effective notification system would need to answer difficult questions about what constitutes a reportable incident, who would make the notification, how quickly information would have to be provided and what information could safely be disclosed. Governments are unlikely to share sensitive technical information if they believe doing so could expose military capabilities, intelligence methods or vulnerabilities.
The system would also need to distinguish between accidents, criminal activity, state-sponsored operations and deliberate military action. Artificial intelligence can be involved in many types of incidents, and identifying responsibility can be difficult even after extensive investigation. A notification mechanism could therefore be useful only if both sides establish sufficient confidence in the information being exchanged.
AI Risks Are Moving Into National Security
The proposed dialogue reflects a broader change in the way governments view artificial intelligence. Earlier debates focused heavily on issues such as employment, misinformation, privacy and regulation. National security agencies are now increasingly concerned about the use of advanced AI in cyber operations, intelligence analysis, military systems and attacks against critical infrastructure.
The United States has already moved to integrate advanced AI systems more deeply into national security activities. A presidential memorandum issued in June directed the US national security establishment to accelerate the adoption of advanced AI while requiring systems used in sensitive applications to be reliable, controllable and subject to clear accountability.
China is also treating artificial intelligence as a strategic technology. Its development of advanced domestic models has continued despite restrictions on access to some high-end American computing technology. The result is a competitive environment in which progress in AI is simultaneously viewed as an economic opportunity and a national security capability.
That makes communication more important but also more complicated. Each side has reasons to understand the other's capabilities while also having reasons to conceal them.
The AI Channel Could Be Separated From Technology Disputes
One of the most notable aspects of the New York discussions is that US Trade Representative Jamieson Greer said export controls on advanced AI chips and semiconductor manufacturing equipment were not part of the proposed AI safety mechanism. This separation could make limited cooperation easier because it avoids immediately linking safety discussions with the most contentious part of the technology relationship.
The distinction could allow Washington to maintain its technology restrictions while still establishing a communication channel with Beijing. China, meanwhile, could participate in discussions about AI risks without accepting that American export controls are legitimate or agreeing to limits on its own technological development.
Such compartmentalisation is already visible in the wider economic relationship. The New York talks also involved plans to operationalise a new trade process covering potentially lower tariffs on selected non-sensitive goods, including consumer products, energy, agricultural goods and medical devices. The discussions therefore appear to be pursuing limited areas of practical cooperation while leaving larger strategic disagreements unresolved.
Trust Remains the Main Constraint
The biggest difficulty is not necessarily technical. It is political trust. Both countries have accused each other of using technology in ways that could threaten national security, and both view technological leadership as closely connected to economic and strategic power.
That makes the meaning of a notification particularly sensitive. If Washington reports an AI-related security incident and Beijing disputes the assessment, the mechanism could become another source of disagreement. Conversely, if China reports an incident involving American technology and Washington rejects the allegation, the dialogue could quickly become entangled with existing disputes.
The fact that an earlier proposal for consultations on AI development discussed by Trump and Xi in May was never formalised illustrates the difficulty of converting political interest into permanent institutional cooperation. The latest proposal therefore represents an opening rather than an established system.
A Limited Mechanism Could Still Have Wider Value
The most realistic value of the proposed arrangement may lie in crisis communication rather than broad AI governance. Establishing a reliable channel between officials could provide a way to communicate when an incident develops faster than conventional diplomatic processes can respond.
Other governments could eventually benefit from such an approach if the United States and China succeed in defining workable procedures. AI systems are increasingly interconnected with global digital infrastructure, meaning a major incident affecting one country's systems could have consequences elsewhere. A bilateral mechanism between the two largest AI powers could therefore provide experience for wider international arrangements.
However, its effectiveness would depend on whether both sides can maintain communication during periods of heightened tension. A notification system that functions only when relations are stable would have limited value. The real test would come during a serious incident when national security officials have incomplete information and strong incentives to assume hostile intent.
The New York talks therefore mark an important change in the AI relationship without resolving its central contradictions. Washington and Beijing remain competitors over advanced technology, computing capacity and strategic influence. At the same time, the growing consequences of AI create risks that neither side can manage entirely through unilateral action.
The proposed notification mechanism addresses that narrow space between competition and cooperation. Its significance will ultimately depend on whether the two governments can turn the idea into a defined process with clear thresholds, reliable communication and enough mutual confidence to function during a crisis. For now, the proposal shows that technological rivalry is creating an unexpected requirement: the two leading AI powers increasingly need mechanisms for communicating precisely because they do not trust each other enough to assume that future AI-related crises will remain contained.
(Source:www.nbcnews.com)
The proposal comes at a time when artificial intelligence has moved beyond being primarily a commercial and technological issue. Advanced systems are increasingly relevant to cybersecurity, military capabilities, intelligence, critical infrastructure and economic competitiveness. That expansion creates a practical problem for both governments: an incident involving artificial intelligence could cross national borders or be interpreted as hostile action before officials have established what actually happened. A direct notification channel could provide a way to reduce that uncertainty.
Competition Is Creating a Need for Communication
Washington and Beijing remain strategic competitors in artificial intelligence, particularly in advanced computing, semiconductor technology and increasingly capable AI systems. The United States has maintained restrictions on exports of advanced AI chips and semiconductor manufacturing equipment to China, while Chinese companies have continued developing domestic alternatives and increasingly capable models. The two countries therefore have strong incentives to protect their technological advantages while limiting risks associated with uncontrolled competition.
That combination makes conventional cooperation difficult. Any discussion about artificial intelligence can quickly overlap with national security and technology restrictions. The latest proposal attempts to separate one narrow area of cooperation from the larger technology rivalry. Instead of seeking agreement on how quickly AI should develop or which country should lead, the proposed mechanism would focus on incidents that could create serious security consequences.
This distinction matters because the two governments do not need to agree on the future of artificial intelligence to recognise that certain incidents could become dangerous through misunderstanding. A major cyberattack involving an AI system, for example, could be interpreted differently by the country experiencing it and the country believed to be responsible. Communication could potentially help establish facts before political or military responses escalate.
The Proposed System Addresses an Information Gap
Bessent described the idea as a notification mechanism covering AI-related incidents that rise to the level of national security. The broader proposal also includes an AI dialogue intended to establish greater transparency between the two leading AI powers. The precise structure, thresholds and obligations of such a mechanism have not yet been publicly defined, and China's public response has so far been limited.
That lack of detail is important. An effective notification system would need to answer difficult questions about what constitutes a reportable incident, who would make the notification, how quickly information would have to be provided and what information could safely be disclosed. Governments are unlikely to share sensitive technical information if they believe doing so could expose military capabilities, intelligence methods or vulnerabilities.
The system would also need to distinguish between accidents, criminal activity, state-sponsored operations and deliberate military action. Artificial intelligence can be involved in many types of incidents, and identifying responsibility can be difficult even after extensive investigation. A notification mechanism could therefore be useful only if both sides establish sufficient confidence in the information being exchanged.
AI Risks Are Moving Into National Security
The proposed dialogue reflects a broader change in the way governments view artificial intelligence. Earlier debates focused heavily on issues such as employment, misinformation, privacy and regulation. National security agencies are now increasingly concerned about the use of advanced AI in cyber operations, intelligence analysis, military systems and attacks against critical infrastructure.
The United States has already moved to integrate advanced AI systems more deeply into national security activities. A presidential memorandum issued in June directed the US national security establishment to accelerate the adoption of advanced AI while requiring systems used in sensitive applications to be reliable, controllable and subject to clear accountability.
China is also treating artificial intelligence as a strategic technology. Its development of advanced domestic models has continued despite restrictions on access to some high-end American computing technology. The result is a competitive environment in which progress in AI is simultaneously viewed as an economic opportunity and a national security capability.
That makes communication more important but also more complicated. Each side has reasons to understand the other's capabilities while also having reasons to conceal them.
The AI Channel Could Be Separated From Technology Disputes
One of the most notable aspects of the New York discussions is that US Trade Representative Jamieson Greer said export controls on advanced AI chips and semiconductor manufacturing equipment were not part of the proposed AI safety mechanism. This separation could make limited cooperation easier because it avoids immediately linking safety discussions with the most contentious part of the technology relationship.
The distinction could allow Washington to maintain its technology restrictions while still establishing a communication channel with Beijing. China, meanwhile, could participate in discussions about AI risks without accepting that American export controls are legitimate or agreeing to limits on its own technological development.
Such compartmentalisation is already visible in the wider economic relationship. The New York talks also involved plans to operationalise a new trade process covering potentially lower tariffs on selected non-sensitive goods, including consumer products, energy, agricultural goods and medical devices. The discussions therefore appear to be pursuing limited areas of practical cooperation while leaving larger strategic disagreements unresolved.
Trust Remains the Main Constraint
The biggest difficulty is not necessarily technical. It is political trust. Both countries have accused each other of using technology in ways that could threaten national security, and both view technological leadership as closely connected to economic and strategic power.
That makes the meaning of a notification particularly sensitive. If Washington reports an AI-related security incident and Beijing disputes the assessment, the mechanism could become another source of disagreement. Conversely, if China reports an incident involving American technology and Washington rejects the allegation, the dialogue could quickly become entangled with existing disputes.
The fact that an earlier proposal for consultations on AI development discussed by Trump and Xi in May was never formalised illustrates the difficulty of converting political interest into permanent institutional cooperation. The latest proposal therefore represents an opening rather than an established system.
A Limited Mechanism Could Still Have Wider Value
The most realistic value of the proposed arrangement may lie in crisis communication rather than broad AI governance. Establishing a reliable channel between officials could provide a way to communicate when an incident develops faster than conventional diplomatic processes can respond.
Other governments could eventually benefit from such an approach if the United States and China succeed in defining workable procedures. AI systems are increasingly interconnected with global digital infrastructure, meaning a major incident affecting one country's systems could have consequences elsewhere. A bilateral mechanism between the two largest AI powers could therefore provide experience for wider international arrangements.
However, its effectiveness would depend on whether both sides can maintain communication during periods of heightened tension. A notification system that functions only when relations are stable would have limited value. The real test would come during a serious incident when national security officials have incomplete information and strong incentives to assume hostile intent.
The New York talks therefore mark an important change in the AI relationship without resolving its central contradictions. Washington and Beijing remain competitors over advanced technology, computing capacity and strategic influence. At the same time, the growing consequences of AI create risks that neither side can manage entirely through unilateral action.
The proposed notification mechanism addresses that narrow space between competition and cooperation. Its significance will ultimately depend on whether the two governments can turn the idea into a defined process with clear thresholds, reliable communication and enough mutual confidence to function during a crisis. For now, the proposal shows that technological rivalry is creating an unexpected requirement: the two leading AI powers increasingly need mechanisms for communicating precisely because they do not trust each other enough to assume that future AI-related crises will remain contained.
(Source:www.nbcnews.com)