Amid rising geopolitical tensions, a new report suggests that collaboration between the United States and China on artificial intelligence (AI) governance and safety remains achievable in certain key areas. Published in January 2025 by a consortium of AI governance experts, the Oxford Martin School policy paper, titled “Promising Topics for US-China Dialogues on AI Safety and Governance” identifies promising topics for dialogue between the world’s two leading AI superpowers.
Building on the Bletchley Declaration of 2023, signed by both countries, the paper underscores a shared recognition of AI’s transformative potential and its associated risks. Despite significant political and economic competition, the document outlines areas where mutual interests converge, particularly in AI safety, risk management, and technical standards.
The paper’s authors advocate for enhancing bilateral discussions on AI safety, focusing on evaluating the “dangerous capabilities” of advanced AI models. Both nations share concerns about preventing the proliferation of AI-enabled capabilities, such as cyberattacks and chemical or biological threats, to non-state actors. The paper suggests that the U.S. and China could agree on critical thresholds and red lines for AI systems, potentially setting a global benchmark for assessing AI-related national security risks.
Technical standards for AI reliability, robustness, and adversarial testing also emerged as a promising area for collaboration. The paper recommends leveraging existing international bodies, such as the International Standards Organization, or establishing new forums to align commercial AI product safety standards. This alignment could facilitate global trade and reduce risks associated with the deployment of unreliable AI systems.
Beyond government-to-government talks, the paper highlights the potential for industry-driven initiatives. For instance, international consortia like the Coalition for Content Provenance and Authenticity could involve Chinese companies in ongoing efforts to develop tools for content traceability and watermarking. These measures address growing concerns about the misuse of generative AI technologies, including the spread of misinformation.
The paper also emphasizes the role of “Track II” dialogues—informal discussions among scientists, academics, and industry leaders—in shaping innovative governance mechanisms. These dialogues could explore novel approaches, such as model registration systems and AI-driven safety evaluations, to enhance global AI oversight. By focusing on non-sensitive topics, Track II engagements could bypass political barriers and foster trust between the two nations.
The paper positions these recommendations within the broader context of U.S.-China relations, which have grown increasingly complex in recent years. While both nations recognize the importance of AI governance, their approaches differ significantly. The U.S. prioritizes democratic values and decentralized regulation, while China emphasizes centralized oversight and content control. Despite these differences, both countries have demonstrated a willingness to cooperate on AI-related issues, as evidenced by their joint support for the United Nations’ 2024 resolution promoting safe and trustworthy AI systems.
However, the paper acknowledges potential risks associated with deeper collaboration. For example, while sharing technical evaluation methodologies might advance AI safety, it could also inadvertently expose sensitive technologies. The authors caution that identifying suitable topics for cooperation requires careful consideration of these trade-offs.
The paper concludes that while geopolitical tensions may hinder some forms of cooperation, there remains significant potential for progress in areas of mutual interest. By focusing on shared goals—such as enhancing AI safety and establishing robust technical standards—the U.S. and China could lead global efforts to mitigate AI’s risks while unlocking its transformative potential.
Need Help?
If you have questions or concerns about any global guidelines, regulations and laws, don’t hesitate to reach out to BABL AI. Their Audit Experts can offer valuable insight, and ensure you’re informed and compliant.
