US and China Seek AI Safety Talks Despite Tech Rivalry
US and Chinese officials propose incident notification mechanisms to mitigate shared AI risks during upcoming high-level talks.
Key points
- US and China propose a notification mechanism for AI incidents affecting national security.
- Both nations identify cyberattacks and infrastructure threats as shared AI risks.
- Geopolitical competition continues to complicate efforts to establish global AI safety standards.
Leaders from the United States and China are preparing to meet this week to address growing concerns over the safety of artificial intelligence. The discussion comes as both nations intensify their competition for technological dominance, creating a complex dynamic where strategic rivalry intersects with a mutual fear of uncontrolled AI systems.
According to reporting from PBS, U.S. Treasury Secretary Scott Bessent indicated that discussions have included the creation of a new notification mechanism. This protocol would allow both countries to alert each other in the event of AI incidents that pose a threat to national security, such as severe cyberattacks or loss of system control.
Shared risks drive diplomatic efforts
Despite the broader geopolitical tensions, experts note that both governments face similar threats regarding the misuse of AI. These risks include the potential for AI-enabled attacks on critical infrastructure, such as power grids, financial systems, and telecommunications. Officials from both sides have acknowledged that these dangers do not respect borders, making cooperation on safety protocols a logical necessity.
Michael Kratsios, the White House science and technology adviser, emphasized the importance of avoiding shared risks as a primary goal for the dialogue. He suggested that focusing on common threats could yield fruitful results, even if other aspects of the relationship remain strained. This approach aims to isolate technical safety from the wider competitive race for AI supremacy.
Competition complicates safety cooperation
However, the path to cooperation is not straightforward. President Donald Trump has repeatedly stated that the United States must maintain a technological edge over China to ensure national security. This stance contrasts with calls from some American tech leaders to slow down AI development to prioritize safety. The tension between winning the race and managing the risks creates a difficult balance for policymakers.
Analysts warn that the U.S.-China conversation on AI is heavily influenced by the broader diplomatic relationship. As China narrows the technology gap, Washington expresses concern about how advanced AI might be used against American interests. This mutual suspicion means that while both sides want to prevent catastrophic failures, they remain wary of sharing sensitive information that could compromise their competitive position.
Building technical trust through dialogue
Experts from Tsinghua University and Johns Hopkins University suggest that establishing clear communication channels for emergencies could help build technical confidence over time. By agreeing on how to report and handle AI-related incidents, both nations can create a framework that reduces the risk of accidental escalation. This step-by-step process is seen as a viable way to manage the technology's dangers without requiring full-scale strategic alignment.
