Written evidence submitted by Oxford China Policy Lab (SDY0019)

Science Diplomacy

Sam Hogg is Head of Policy Engagement of OCPL. His work focuses on geopolitics, AI and strategy. Previously, he founded Beijing to Britain, a company that produced and delivered research notes on the UK-China bilateral weekly to thousands of Government readers, politicians, embassies and corporate clients. Read a million times across 140 countries, various governments, the FTSE100, and embassies around the world, his writing on strategic thinking in Government and the need to build China capabilities has been published on multiple occasions by the British Parliament.

Kayla Blomquist is Director and co-founder of the Oxford China Policy Lab (OCPL) and a DPhil researcher at the Oxford Internet Institute, with an affiliation to the Oxford Martin School AI Governance Initiative. Her research focuses on US-China relations and international AI governance, examining how status and legitimacy concerns shape China and US AI governance and development. Prior to Oxford, she served as a diplomat in the U.S. Mission to China for four years, specialising in governance of emerging and dual-use technologies. She is professionally fluent in Mandarin Chinese. Her insights and research have been shared globally in media outlets including The Economist, BBC, ABC, and Reuters.

While this inquiry addresses a broad landscape of UK science and innovation, this submission will focus on one high-priority and fast-evolving domain: artificial intelligence. Specifically, we will examine the UK’s current landscape, and how it can strengthen its international positioning through excellence in AI governance, safety, and security.

To what extent is the UK considered a global leader in science and innovation and how does this contribute to its soft power on the global stage?

1. The UK has developed substantial soft power through its leadership in AI and other emerging technologies, particularly in areas of governance, safety, and security. In simple terms, the UK has sought to present itself as a ‘norm entrepreneur’ on AI safety and security issues, with a hope that other nations will follow or copy its efforts. As this evidence explores, the UK’s  soft power, defined as the ability to get what you want through attraction rather than coercion, allows it to project its AI aims abroad with credibility.[1]

2. The image Britain wants to convey domestically and internationally is one of leadership in science and technology, with a particularly strong AI research and development (R&D) landscape and a commitment to working with partners to harness these technologies safely. By several measures, the UK “outperforms [its] closest competitors” and stands as a “main challenger” to the United States and China in key domains, while lagging behind in AI frontier firms and end-to-end development capability.[2]

3. For example, the UK boasts four of the world’s top 10 universities and a technology sector worth over $1 trillion. It produces a disproportionate share of high-impact research - ranked third globally for highly cited outputs - reflecting scientific excellence that far exceeds the UK’s size.[3] Breakthroughs originating in the UK, including DeepMind’s advances in AI, underscore a long history of innovation contributing to global progress.[4] The UK has already begun to develop strategic guidance on the diffusion of AI throughout its economy and society,  and commissioned extensive research on the potential impacts of AI on defence and security, leading the majority of other countries.[5][6]

4. To this end, the UK has aimed to convey its AI norms and views credibly, and in doing so build its soft power, through the following actions:

1. Convening and creating: The Government launched the world’s first AI Safety Summit and created an AI Safety (now AI Security) Institute (AISI). Other nations have since launched their own versions of AISIs and continued the AI Summit series. At the inaugural AI Safety Summit, the UK was able to bring two rival AI superpowers to the table - China and the United States - during a time of geopolitical tension, along with 27 other countries. The UK goes beyond government-to-government engagement, and has drawn attendance at key events from more than a dozen leading AI and technology companies, including Anthropic, Alibaba, Google, Nvidia and XAI.[7] Many of these companies engage with the AISI.

2. High level strategy and communications on safe and secure AI: Senior British politicians have spoken about AI and the UK’s approach to the technology for years, covering topics from how to utilise AI in the economy to dealing with security and risk related issues which AI can expose and enhance. In recent months, it has added warnings around issues such as Artificial General Intelligence (AGI) into their speeches (see examples below). This consistent drum beat of rhetoric backed by concrete action and strategic norm creation and promotion makes the UK’s AI leadership attractive to other nations (evidenced in part by how frequently it is raised in diplomatic meetings, agreements and readouts) and helps build the UK’s credibility and prestige.

3. Top level engagement: The UK continues to engage internationally with other nations - from close partners like the United States or those in the Five Eyes, to more challenging nations like China - on AI opportunities, risks, and norms. The AISI spells out some of those risks in its recent ‘Research Agenda’ note, but these include issues such as cyber misuse, criminal misuse, societal resilience and human influence.[8]

Convening and launching

5. At a very high level, the UK’s soft power has already been demonstrated through exceptional diplomatic prowess by bringing together the US and China, the world's two leading AI powers, during heightened geopolitical tension at the 2023 Bletchley Park AI Safety Summit, and in the number of other nations that launched their own versions of the AISI.

6. This is because Britain occupies an influential position in the global AI ecosystem as a norm-shaper, security convener, and innovation hub. For example, the UK AISI has pioneered critical norms in international AI governance that have substantially advanced both the UK's national security interests and its soft power, building its credibility.[9][10]

7. To give one tangible policy example, the AISI has an area of focus on “the use of AI to make child sexual abuse images”, with a new team exploring methods to help to prevent abusers from harnessing the technology to violate the law.[11] This problem is not constrained by national borders, and methodological advances made by the AISI can be shared internationally. As a second example, the UK consulted globally to develop its Code of Practice for the Cyber Security of AI.[12] As Science, Innovation and Technology Secretary Peter Kyle recently noted in a speech on AI, “a security risk is a security risk, no matter where it comes from.”[13]

8. Zooming out, the Government (the Ministry of Defence’s Defence AI & Autonomy Unit and the Foreign, Commonwealth and Development Office) jointly commissioned a short scoping study from RAND Europe on ways in which military use of AI might generate risks and opportunities at the strategic level. The resulting research was shared publicly and covered risks around issues such as AGI and nonalignment through to information-manipulation (e.g. deepfakes), helping project a forward-looking approach.[14] Similarly, Pat McFadden, Chancellor of the Duchy of Lancaster, delivered a speech on AI and cyber.[15] He made a point of declassifying an intelligence assessment on AI’s potential impact on cyber through to 2027, produced by the NCSC.[16]

9. Since its formation, the UK’s AISI has secured first-of-their-kind commitments from frontier AI companies to develop safety frameworks to protect against the most extreme harms of AI and attracted some of the world’s most cutting-edge AI research talent.[17][18] DSIT Secretary Kyle acknowledged this in a speech at the Munich Security Conference earlier this year “[AI risks] aren’t distant possibilities. They are real, tangible harms, happening right now. The implications for our people could be pervasive and profound. In the UK, we have built the largest team in a government dedicated to understanding AI capabilities and risks in the world. That work is rooted in the strength of our partnerships with the companies who are right at the frontier of AI.”[19] More widely, the British Government has commissioned and shared research on long-term thinking on AI;  published an AI action plan; and regularly publishes updates on roll-outs of AI in different parts of society.[20]

10. Just as countries have historically led through setting global standards, the UK has an opportunity to lead in AI not by dominating model development, but by becoming the trusted benchmark for responsible innovation: the ‘quality setter’ for a global AI ecosystem. By continuing to commission research through the AISI or wider government, and release it publicly or discuss it with trusted partners, it has solid footing to execute this approach.

High-level strategy and communications on safe and secure AI

11. AI safety, security and opportunity has been a particular point of continuity from the Conservative Government of then-Prime Minister Rishi Sunak to now Labour Prime Minister Keir Starmer, who launched Britain’s first AI Opportunities Action Plan, and has further empowered his Cabinet to speak on AI issues in a global context.[21][22] More generally, the Government has explicitly committed to “taking action to ensure that we can reap the benefits of AI while mitigating potential risks and harms. This includes acting to establish the right guardrails for AI through our agile approach to regulation; leading the world on AI safety by establishing the first state-backed organisation focused on advanced AI safety for the public interest; and – since 2021 – encouraging the development of a flourishing AI assurance ecosystem.”[23]

12. The Government inserts AI analysis into critical, internationally-facing speeches, on a number of different fronts:

13. For example, in the first major British diplomatic speech on the issue, Foreign Secretary David Lammy discussed AGI, noting that "As we move from an era of AI towards Artificial General Intelligence, we face a period where international norms and rules will have to adapt rapidly. It is absolutely essential that the UK, US and our allies’ governments and companies work together ever more closely to win that technological race.”[24]

14. DSIT Secretary Kyle discussed wider international engagement on AI at the heavily scrutinised Munich Security Conference, saying that“the AI revolution is happening. Ignoring it is simply not an option. In the UK, we reject the doomsayers and the pessimists…Working in collaboration with our international partners, we’re going to create one of the biggest clusters of AI innovation in the world and deliver a new era of prosperity and wealth creation for our country.”

15. Pat McFadden, the Chancellor of the Duchy of Lancaster, used a speech to focus on AI and cyber scenarios in the coming years, noting “The stakes are high right now. And we are in the middle of another huge technological leap - a “technology shock” if you like - with AI and other emerging technologies developing at breakneck speeds...AI is going to increase not only the frequency, but the intensity, of cyber attacks in the coming years. Our security systems will only remain secure if they keep pace with what our adversaries are doing.[25]

Top-level engagement

16. This early AI leadership allows Britain to shape the strategic landscape to serve its interests, creating a more secure approach to both national and international technology development. Senior leaders from overseas political and private sector backgrounds attended the Bletchley Park gathering, including from many of the major frontier AI companies which sit beyond the UK’s immediate jurisdiction.[26]

17. Evidence would suggest that this is received well internationally: for example, many of Britain’s recent bilateral discussions with countries around the world reference AI partnerships or plans, including recent talks with the United States, AI partnerships with Qatar, and emerging technology agreements with India.[27][28][29]Additionally, research has shown that the UK’s initial AI approach has inspired similar domestic efforts to facilitate industry commitments in China.[30]

How has this agenda been impacted by the current geopolitical environment, including the international activities of Russia and China? 

18. The ever-changing global geopolitical environment has, naturally, impacted and shaped Britain’s own approach and projections on AI. It has cemented the importance in the minds of senior leaders of ‘winning’ the technological contest with China, the need to bolster cyber defence against attacks from hostile nations such as Russia, and, as DeepSeek has shown, to strategise as much as possible around the release of models from challenger nations.[31][32]

19. In his speech at the Munich Security Conference, Mr Kyle specifically highlighted DeepSeek’s release of an open weight frontier model as a critical moment for UK AI strategy, saying “with DeepSeek, we’ve just seen once again just how sudden, how unpredictable, innovation can be. The AI revolution is happening.”[33][34]

20. Many countries are competing to become global leaders in AI, including China, and want to project an image that their methods and approaches are credible.[35] For example, China actively promotes its model of AI governance to countries throughout the world as more forward-leaning and effective than Western approaches. This model integrates advanced surveillance technologies and centralized data collection with the promise of enhanced public safety and economic efficiency.[36]

21. More widely, the British Government holds the view that it and its allies must be the first to “win the race” towards AGI, although experts and analysts have a range of views across the spectrum as to this approach.  Both Mr Lammy and Mr Kyle have also referenced AGI as another key issue requiring action, with fears that “losing oversight and control of [AGI]... would be catastrophic.” Britain’s Ambassador to the United States, Peter Mandelson has written that both nations’ “mission is crystal clear; to prevail in the technological contest with China.”[37][38][39]

22. On other AI-linked issues in the geopolitical landscape, post-Russia’s illegal invasion of Ukraine, the the UK is investing £1.1 million to give the Moldovan Government tools to protect the country’s upcoming Parliamentary Election, and further invested £8 million in the Ukraine Cyber Programme over the next year "to continue to counter the Kremlin’s cyber aggression."[40]

23. The AISI’s first ever research agenda, recently published, prioritises keeping “partners within the rest of the UK government, the US government, and national security partners, as well as…broadly…the Network of AISIs and many international governments” abreast on the state of frontier AI safety.[41]

24. The geopolitical landscape continues to evolve. Some AI companies are using this as an opportunity to push for ‘democratic blocs’ to be created for them to disseminate their products within, whilst blocs such as the African Union (which contains over a dozen Commonwealth countries) or nations such as the United Arab Emirates and Saudi Arabia are strategising around their own positioning in the value chain.[42][43][44] For the UK, this means  its AI strategy and approach needs to balance caution with pragmatism, with a deep understanding of the incentives at play. On a geopolitical level, it can be broken down on a country-by-country basis:[45]
     
1. United States: The US-UK science relationship remains foundational, with deep research collaboration, industry partnerships, and shared approaches to many AI governance challenges.
2. European Union: Prime Minister Keir Starmer has made closer defence and economic engagement with the EU a priority, including scientific collaboration through programs like Horizon Europe.[46]
3. Australia and Canada: Key "Five Eyes" partners with whom the UK shares intelligence and increasingly coordinates on technology security issues. Under Pillar II of AUKUS, the UK already works with Australia on emerging technologies and AI matters.[47] Meanwhile, the UK and Canada have a coherent science of AI safety partnership.[48]
4. Japan: A crucial partner for the UK in advanced technologies, particularly in areas like semiconductor supply chains, quantum computing, and AI. Japan shares many of the UK's concerns about technology security and ethical development, and both nations recently signed a Memorandum of Cooperation on quantum science and technology.[49]
5. South Korea: An increasingly important partner, particularly in AI governance (as demonstrated by the Seoul Summit following the UK's Bletchley Park initiative) and advanced manufacturing. Both nations signed the Downing Street Accord in November 2023, which placed a heavy emphasis on science and technology.[50] In particular, it binds both countries to “recognise the need for global norms and technical standards that promote innovation in digital technology, including artificial intelligence, while promoting our shared values and minimising risks. Building on this shared understanding, we intend to continue discussions on matters of common interest, involving norms and technical standards including advancing AI safety, closing the digital divide, addressing digital literacy, providing citizen-centric digital government services, creating a sustainable digital ecosystem.”[51]
6. China: As the UK's most significant strategic challenger in science and technology, the relationship with China requires complex and strategic national security management. It combines competitive aspects, particularly in frontier technology development, with necessary cooperation on global challenges, security, and governance. As noted, the UK successfully brought China into the AI Safety Summit at Bletchley Park, demonstrating its ability to engage China constructively on technology governance despite tensions. Britain’s engagement with China on AI dialogue is ongoing.[52]

How does science and technology innovation contribute to the UK’s national security? 

25. Scientific and technological innovation are integral to the UK’s national security in multiple dimensions, as noted in both the UK’s most recent foreign policy documents, the Integrated Review, and the Integrated Review Refresh.[53] Cutting-edge capabilities in areas such as AI, cyber, quantum computing, biotechnology, and space not only drive economic prosperity but also provide strategic and tactical advantages that safeguard the nation.

26. For example, the UK’s NCSC has stated AI will “almost certainly continue to make elements of cyber intrusion operations more effective and efficient, leading to an increase in frequency and intensity of cyber threats.”[54] This matters, because the NCSC assesses that Russia plans to run malicious cyber campaigns beyond any ceasefire agreement made.[55]

27. To this end, the AISI works with the Laboratory for AI Security Research (LASR), the National Cyber Security Centre (NCSC), the Defence Science and Technology Laboratory and the national security community on national security and interlinked issues.[56] LASR is tasked with “accelerat[ing] innovation and research into how AI can protect our national security” and has already funded 10 PhDs at Oxford University, and an in-house team of nine researchers at The Turing Institute.[57]

28. AI is integrated on other national security projects, from Project Tempest, the UK's initiative to develop a sixth-generation fighter jet, to how GCHQ thinks about using the technology to counter hostile state-created disinformation.[58][59]The general purpose nature of this technology means AI can be utilised every minute of every day in the UK’s national security infrastructure.

29. Ultimately, as evidenced by both the rhetoric - “AI is already revolutionising many parts of life - including national security" - and the sustained investment on matters from quantum to AI, there is almost no area of the UK’s overarching national security landscape that is not touched by this technology.[60]But AI in the national security space must be strategised and deployed safely, and it should be noted that the UK has made a point of communicating and actioning the integration of AI into the national security space in this  manner.[61]

9th May 2025

9