Bresnick was interested in understanding what Chinese experts themselves were saying about AI capabilities and risks, as much of the Western press coverage focused on China's centralized governance advantage without delving into Chinese perspectives.
China sees AI as foundational for its military modernization, enabling what they call 'intelligentized warfare,' characterized by increased speed, precision, and autonomy. They believe the country that develops these systems faster will have a significant advantage in future wars.
Chinese experts are concerned about the lack of explainability and trustworthiness of AI systems, which could lead to escalations, accidents, or mistakes. They also worry about the potential targeting of friendly forces or civilians by AI systems.
One debate is whether China should focus on high-quality, 'exquisite' AI systems or rely on mass production of 'good enough' systems. Another involves whether the hierarchical Chinese military culture allows for effective use of AI, requiring potential reforms in decision-making processes.
Evidence includes reports in Chinese media, procurement documents, and academic publications discussing specific AI algorithms and military problems. The Department of Defense also highlights Chinese advancements in AI in its annual reports on China's military.
Chinese experts struggle with the 'black box' nature of AI, where decisions are not easily explainable. Additionally, the hierarchical and bureaucratic nature of the Chinese military, where trust is low between superiors and subordinates, further complicates the adoption of AI systems.
Constraints include data issues, such as a lack of militarily relevant data due to the absence of recent conflicts, difficulties in sharing data across military branches, and challenges in digitizing data. Other issues include cybersecurity, testing and evaluation, and the need for standardized AI development.
Chinese experts suggest using synthetic data, combining human knowledge with AI systems, and improving model training to enhance explainability and trustworthiness. They also advocate for international engagement on AI safety, particularly with the United States.
The discussions suggest that AI could change China's approach to deterrence in the South and East China Seas, as the speed and autonomy of AI systems may increase the risk of escalation. There is also a recognition of the need for international dialogue to manage AI-related risks.
Potential areas for discussion include escalation dynamics, norms for testing and evaluating AI-enabled military systems, and the role of AI in cyber defense or space operations. However, some of these topics, like cyber and space, may be too sensitive for immediate engagement.
In China’s 14th Five-Year Plan that spans from 2021 to 2025, priority was assigned to development of emerging technologies that could be both disruptive and foundational for the future. China is now a global leader in AI technology and is poised to overtake the West and become the world leader in AI in the years ahead. Importantly, there is growing evidence that AI-enabled military capabilities are becoming increasingly central to Chinese military concepts for fighting future wars.
A recently released report provides insights on Chinese perspectives on military use of AI. Published by Georgetown’s Center for Security and Emerging Technology (CSET), the report illustrates some of the key challenges Chinese defense experts have identified in developing and fielding AI-related technologies and capabilities.
Host Bonnie Glaser is joined by the author of this report, Sam Bresnick, who is a Research Fellow at Georgetown’s CSET focusing on AI applications and Chinese technology policy.
TimestampsB
[00:00] Start
[01:33] Impetus for the Georgetown CSET Report
[03:34] China’s Assessment of the Impacts of AI and Emerging Technologies
[06:32] Areas of Debate Among Chinese Scholars
[09:39] Evidence of Progress in the Military Application of AI
[12:13] Lack of Trust Amongst Chinese Experts in Existing Technologies
[14:25] Constraints in the Development and Implementation of AI
[18:20] Chinese Expert Recommendations for Mitigating AI Risk
[23:01] Implications Taken from Discussions on AI Risk
[25:14] US-China Areas of Discussion on the Military Use of AI
[28:50] Unilateral Steps Toward Risk Mitigation