Remarks on UK-China AI Relations:

House of Lords Panel on The Future of UK-China Foreign Policy

Kayla Blomquist

27 June 2023

Event record: https://www.thecfhk.org/post/british-parliament-hosts-session-in-house-of-lords-looking-at-future-of-uk-china-foreign-policy 

Introduction:

The United Kingdom faces crucial and time-sensitive considerations in shaping its relationship with the People’s Republic of China (PRC), particularly concerning the advancements and implications of –as well as coordination and competition surrounding– artificial intelligence (AI). While the PRC’s currently known capabilities in AI, especially in large language models (LLMs), currently lag behind those of the US and the UK, the PRC will remain a major player in AI for years to come.^[1]

To navigate the complex and uncertain landscape of geopolitical competition amidst the rapid development and deployment of AI, we must first understand and acknowledge several realities.

First, current policy and media discourse prominently frames the PRC as a threat and technological rival. Indeed, the PRC’s deployment of AI-related technologies for human rights abuses^[2] and the People's Liberation Army's pursuit of global military superiority through AI development^[3] pose specific challenges and threats to the UK’s values and strategic interests. Second, however, it is also vital to recognize the potential benefits of joint research, scientific cooperation, and investment with actors within China. For example, the UK’s world-leading academic and research institutions depend on and benefit immensely from the talent and contributions of PRC national students and researchers. Finally, AI will present previously unseen risks to global society which will impact countries no matter their system of governance, location, or level of development. Solutions to these challenges will require effective coordination between the UK, its Western allies, and the PRC; failure in coordination may significantly amplify such risks.

These seemingly conflicting realities are not inherently at odds with one another. Sound, mature policy strategies at the intersection of AI and UK-China relations will take all three into account. Below, I suggest four categories of recommended policies that will empower the UK to a) mitigate technology transfer risks with the PRC in AI, b) create needed space for coordination with the PRC on common global interests in relation to AI, and c) grow into a flourishing, legitimate, and much-needed global leader in AI. This multi-pronged approach will foster an expertise- and value-driven, rather than fear-based, comprehensive policy strategy.

Policy Recommendations

1. Enhance HMG's In-House Subject Matter Expertise at the Intersection of China and AI:

HMG must invest substantial resources in bolstering the UK government's expertise and capability to effectively address the multifaceted issues that sit at the intersection of AI and UK-China relations. This includes investing in research, education, and training programs to develop a skilled internal workforce capable of handling this complex set of questions. Additionally, it requires the creation of more roles for both China- and AI-literate policy advisors and policy implementers.

Recently, the Oxford China Policy Lab and Beijing to Britain co-authored Parliamentary evidence that called for a systematic increase in HMG’s China capabilities.^[4] These recommendations were heeded in the 2023 Integrated Review Refresh, where the UK’s funding for China capabilities was doubled.^[5] The need for government capabilities in emerging technologies was also acknowledged. Since the IR Refresh, multiple technology-focused China analyst positions have been created, marking definite progress in this direction.

However, these actions need to be taken further so that technological and China-facing capabilities are not developed in silos. Rather, this intersectional expertise must be cultivated and incorporated throughout HMG, in positions such as vetting for dual-use technology transfer through research partnerships or business mergers and acquisitions.^[6]  Each of the below policy recommendations will only be possible with this intersectional expertise-based approach.

2. Safeguard Against the Transfer of Dual-Use AI Technologies:

HMG should continue to research and recognize the potential risks associated with increasingly advanced forms of AI and take measures to prevent their unauthorized transfer, particularly to PRC military actors. This will require effectively implemented export controls over advanced semiconductors and, potentially down the road, highly advanced AI models, to the extent possible. The mechanisms available and most critical approaches to this require further investigation, with high priority. Additionally, further caution is warranted in research partnerships, acquisitions, and collaborations surrounding advanced forms of AI.

Specifical actions should include:

• A holistic investigation into high-risk scenarios in dual-use technology transfer including but extending beyond traditional export controls on products.
• An investigation into loopholes in the existing export control regime that might still allow for the discrete transfer of high-end semiconductors to the PRC for military end-use.
• Finally, to make possible future arms control verification regimes surrounding highly advanced AI systems, investment in developing technical verification techniques and corresponding regulatory capabilities.

3. Lay a Constructive Foundation for International Coordination on Global Risks from AI:

The UK is well-positioned to step into a leadership position where it creates space for tangible international coordination to address global risks arising from AI, particularly as a middle actor amidst rocky US-China relations. The UK should engage with other nations, including the PRC, to develop shared frameworks, protocols, and risk management approaches at various levels of government. Perhaps most importantly, it should preserve and further cultivate diplomatic lines of communication that will serve as a foundation for coordination with the PRC when mitigating the risks from highly advanced forms of AI.

Doing so necessitates managing political and diplomatic rhetoric carefully, striking a balance between holding the PRC accountable for unacceptable behavior (such as the illicit transfer of sensitive technologies) and avoiding unnecessary and counterproductive labels of Western coordination in AI as being motivated by excluding the PRC. While excluding the PRC from certain fora may be necessary due to safety concerns or natural when coordinating amongst allies, it is equally important to minimize unnecessary antagonism in the framing of these efforts and identify areas where collaboration is desirable and feasible. These considerations are critical as the UK prepares to host its global summit on Artificial Intelligence later this year.^[7] 

By actively cultivating collaboration and constructive dialogue, the UK can contribute to global efforts in addressing AI challenges. The world is in need of a true leader in this respect.

4. Foster a Domestic AI Ecosystem Which Allows the UK’s Democratic and Open Society to Flourish:

In order to grow into a legitimate global leader in AI, the UK must cultivate a holistic AI development and governance ecosystem that nurtures a flourishing democratic system amidst the transformative impacts of AI. This is critical to develop in the face of the PRC’s active and now long-standing cultivation of its own domestic model of AI governance, which it promotes to countries throughout the world as more forward-leaning and effective than current Western approaches.^[8] 

This will be no simple feat. For example, there is no straightforward solution to the inherent tensions between the core value of freedom of speech and content moderation in generative AI systems and online platform management. And, needed regulations mitigating the harmful effects of emergent forms of AI may inherently inhibit much-desired innovation and industry growth. A series of delicate balances will need to be found and negotiated by society over time.

In navigating these challenges, the UK must continue to celebrate and invest in its democratic traditions. First, it must build into its system of AI development and governance the robust engagement of civil society, drawing upon its world-class expertise at home. It should further establish mechanisms for auditing emergent AI models before and after their release to assess and mitigate the wide range of risks that they may pose to both UK and global society. Additionally, the UK should invest in applications of AI to safeguard and improve its democratic process, as well as enhance the quality of its public services. Finally, the UK government should directly fund additional policy and academic research programs that develop comprehensive, multidisciplinary approaches to advancing the UK’s model of AI governance.

In Conclusion:

Expanding the UK’s current definition of AI leadership is imperative. It must include not only the development of the most advanced AI models but also the integration of these technologies into society in a way that furthers rather than undermines the UK’s fundamental democratic values. Additionally, the UK must work to lead diplomatically, creating avenues for legitimate and productive forms of international coordination with the PRC for mitigating global risks that AI will pose.

By following the above policy recommendations, the UK will be better equipped to navigate the challenges and opportunities presented by AI, as well as the surrounding geopolitical considerations with the PRC.

This multi-pronged, expertise-driven approach is key if the UK seeks to position itself as a global leader in shaping a responsible and democratic AI-powered future.

Thank you.