Introduction

Henry Kissinger once wrote: ‘Every international order must sooner or later face the impact of two tendencies challenging its cohesion: either a redefinition of legitimacy or a significant shift in the balance of power’ (Kissinger 2015). With the advent of Artificial Intelligence (AI), the central role that digital technology and especially generative Artificial Intelligence is coming into play in geopolitics and international order is undeniable, however, the way in which and the extent to which generative Artificial Intelligence influences and transforms the relationship between two of the world’s most dynamic digital powers—the European Union and China—is still unknown territory. The EU and China are both regarded as leaders in developing and deploying generative AI technologies, yet both have taken divergent approaches to governing AI. Sino-European tensions have mounted given the differences between the EU and China regarding geopolitical crises including the conflict in Ukraine, the war in Gaza and the ongoing ‘Chip War’ between China and the West at large. Amid such polycrises, the risks and opportunities that the proliferation of Generative AI technology holds for international politics could amount to further escalation, and may seriously harm diplomatic ties between the EU and China.

This article explores the ways in which Generative AI impacts diplomatic relationship-building through the prism of EU–China relations. The first section examines the European and Chinese models of Generative AI. It underscores that both the EU and China harbour the ambition to become a global geopolitical power leveraging technologies such as generative Artificial Intelligence and are increasingly prepared to weave technology priorities into the fabric of international diplomacy and geopolitical evolution. The second and third sections investigate the promises and pitfalls of generative AI in EU–China diplomatic relationship-building. The fourth section investigates the potentiality that generative AI could bring to generate unforeseen enabling forces for EU–China diplomatic relationship-building. This paper purports that the distinctive EU and Chinese models of generative AI too often belie the opportunities that could potentially enable the EU and China to build their relationship, since generative AI raises shared concerns for the EU and China, and utilising generative AI could make their communications become more efficient, the EU and China may come to reach some kind of shared framework for generative AI development and governance; this could lead to productive talks in other domains such as the trade deficit issue plaguing the EU–China relations or more sensitive cross-strait issues; moreover, in the realm of public diplomacy generative AI could facilitate the EU and China’s public diplomatic efforts towards each other.

The European and Chinese models of generative AI

‘AI is not just a technical matter or just about intelligence; […] AI is political through and through’ (Coeckelbergh 2022). Political issues such as power and governance take on new urgency and meaning in light of technological developments such as AI. In an era of hybrid diplomacy (Bjola and Manor 2022), AI has become a new tool for diplomacy: international actors with advanced AI can exercise more political, military, economic and diplomatic power. Countries around the globe are seeking to harness AI’s potential for economic advantage, technological supremacy and influence over global norms and standards; in particular, the proliferation of generative artificial intelligence systems raises a myriad of economic, political and ethical issues for diplomatic relationship-building. While technonationalism often dominates the conversation regarding the potential impact of AI on international relations (Ding 2022), there is a dearth of understanding and comparison of the aims, best practices and outcomes of China’s and the EU’s approaches to AI (Roberts et al. 2023), moreover less is understood about how generative AI advances could influence EU–China relations.

Generative AI refers to the ‘AI systems whose main function is to create new content, from images and text to videos and data. Its applications are based on foundation models—AI systems trained on a large quantity of data that can be used for a wide range of downstream tasks’ (Feldstein 2023). Two distinct ways of approaching generative AI are now emerging amongst this AI-driven era of international politics—the European model and the Chinese model. Anchored in the EU’s search for digital strategic autonomy, AI regulation has become a ‘central policy question’ in the European Union (EU). EU policy-makers pledged to develop a ‘human-centric’ approach to generative AI to ensure that Europeans can ‘benefit from new technologies developed and functioning according to the EU’s values and principles’ (European Commission 2019). The EU has been pursing a strategy of digital innovation over the past 20 years, the European Commission established the Digital Single Market strategy in 2015, laying the groundwork for shaping the digital transformation through the Digital Single Market strategy of 2015–2019 and its array of policies and funds to build a foundation for a digital economy and society for all Europeans. In 2018, the European Commission published a European Strategy for AI (European Commission, COM (2018b) 237) and a Coordinated Plan on Artificial Intelligence, setting out ambition to become the ‘world-leading region for developing and deploying cutting-edge, ethical and secure AI, promoting a human-centric approach in the global context’ (European Commission, COM (2018a) 795). This EU model of generative artificial intelligence was further strengthened by the release of the EU’s first White Paper on Artificial Intelligence: a European approach to excellence and trust in 2020, and the Commission proposal for the EU AI Act in 2021, which would be the world’s first comprehensive AI law, albeit currently under negotiation for ratification at the European Parliament. Furthermore, the EU is aiming to build a new global AI framework built on three pillars: guardrails, governance and guiding innovation, to ensure the use and governance of generative AI are grounded in democratic values—a strategy that fits into the EU’s broader diplomatic framework of being a global norm entrepreneur.

However, the China model of regulating generative AI is more government or state controlled. China sees technologies such as AI of strategic importance for the country to strengthen its ‘comprehensive national power’. The importance of AI technologies to China’s modernisation dates back to at least 2015, when the State Council launched the ‘Made in China 2025’ strategy to transform China into a major global player in high-tech manufacturing. The ‘Made in China 2025’ strategy published in 2015 articulates how China aspires to become a world leader in high-tech industries, strengthening domestic innovation and reducing its reliance on foreign technologies, including AI, while moving up in global value chains. Since then, AI industry has been deemed a priority. In 2016, another policy—the ‘“Internet+” AI Three Years Implementation Plan’, was jointly issued by the National Development and Reform Commission, the Ministry of Science and Technology, the Ministry of Industry and Information Technology and the Cyberspace Administration of China. In 2017, China’s State Council released the country’s national AI strategy entitled the ‘New Generation AI Development Plan’, known as the AIDP, setting out China’s ambition of become a leading AI power by 2030. In July 2023, the Cyberspace Administration of China (CAC) released the ‘Interim Measures for the Management of Generative Artificial Intelligence Services’, placing a strong emphasis on the development and deployment of generative AI, showcasing China’s aims to foster innovation and responsible use of generative AI in various sectors, targeting specific applications or manifestations of a technology, centring on specific algorithms, making algorithms the fundamental unit for transparency and disclosure via the algorithm registry, which is different from the European Union’s AI Act.

Contradictions between the two models

The Chinese model is anchored in China’s way of promoting ‘internet sovereignty’ (Zeng et al. 2017) and China’s pursuit of technological superiority on the global stage and utilising non-democratic measures such as mass surveillance for domestic governance. While China has been actively promoting the development and adoption of generative AI, the government has also been actively using generative AI for surveillance, censorship and security control. The Chinese government has been actively supporting the development of local AI. AI’s application in state governance is one of the distinct aspects of China’s AI policy. Chinese local governments are competing to be pioneers of ‘intelligent government’. For example, Guangzhou municipal government introduced facial recognition and Zhejiang provincial government has utilised AI and machine learning to process provincial data to establish a comprehensive provincial government affairs knowledge base (Zeng 2020).

In contrast, the EU’s approach to developing generative AI differs from China’s and centres first and foremost on ‘harnessing its regulatory and norm establishing power to define clear boundaries for the ethically acceptable development and use of AI’ (Csernatoni 2019). The EU’s model is based on democratic values and guarding against societal challenges such as disinformation; the Commission has been prioritising tackling the crucial ethical challenge of the generative AI era which lies in the fact that generative AI exacerbates tensions around data sovereignty, intellectual property rights and technological dependencies.

In parallel with the rapid development of the European and Chinese models of generative AI is AI’s increasing ability to exacerbate the scope and scale of disinformation with ‘deepfake’ technology becoming increasingly accessible and sophisticated in their ability to mimic human interaction (Arsenault and Kreps 2022). These contradictions could exacerbate differences and breed tensions between the EU and China. The relationship between government authorities and AI companies in developing and applying AI is another conundrum that the EU and China face: information about how and when generative AI systems retain and use information entered into them is limited, there is an increasing consciousness in government agencies to ensure that generative AI is developed and deployed responsibly and safely. At the world’s first AI Safety Summit held in the UK in November 2023, leading artificial intelligence companies have agreed to allow government agencies, including those in the EU, to test the safety of new AI models before being released.

Mutual concern

Although generative AI raises problems that already divide China and the EU like intellectual property rights, tech theft as well as ethical issues that EU values more than China, the differences between the EU and Chinese models of generative AI evaluated through a narrow frame focussing on the pitfalls that the proliferation of generative AI hold on liberal democracy and authoritarianism, risk neglecting the potentiality of AI technology being an enabling force in diplomatic relationship-building and an accelerating force that could produce unforeseen transformations in the relationship between the EU and China.

This is largely because the emergence of generative AI equally raises issues of mutual concern for both the EU and China: just as the EU fears the use of generative AI to spread disinformation, swing public opinion and polarise publics so the Chinese fear the use of generative AI to undermine Party Rule and penetrate the Great Fire Wall. Additionally, both the EU and China have shared concerns about the issues of data sovereignty and tech dependence as neither wants to be dependent on the US or have a US-dominated global market. Moreover, as generative AI systems grow increasingly capable of generating realistic text, images and videos, the digital evolution raises considerable ethical issues, with the raise of highly targeted deepfake; generative AI creates direct diplomatic and political risk through its ability to produce disinformation campaigns at scale, that undermines trust in diplomatic institutions, which would be crucial for the EU and China to address jointly. So, counter-intuitively, generative AI includes areas where the EU and China may have shared concerns that could and should align, and both sides could join forces to explore how best to harness the benefits of AI while managing its potential perils, which could be a rationale for severing tensions between the EU and China.

EU–China dialogue in the generative AI era

How can the power conferred by generative AI be utilised to create opportunities for EU–China cooperation? Firstly, generative AI could help accelerate EU–Chinese negotiations, and increase the efficiency of meetings, thanks to automatic translation. With generative AI’s ability to synthesise vast amounts of information, it has the potential to increase decision-making efficiency and improve diplomatic coordination. Generative AI has the unique capability to expedite international negotiations, which is especially pertinent for facilitating diplomatic relationship-building between the EU and China. Cultural factors affect international negotiations (Cohen 1997). Translation which is time consuming have inhibited EU–China negotiations. EU–China negotiations often are scheduled for 3 h, due to simultaneous translation, effective discussions that may generate substantial outcomes only take place for 45 min (Interview with diplomats 2022). Trade negotiations between the EU and China are often the longest in the EU’s external negotiations (Interview with EU Trade Negotiator 2023). The prolonged 7 years of negotiations between the EU and China that led to the ambitious and contentious Comprehensive Agreement on Investment (CAI) is testimony to the linguistic and cultural barriers in negotiations between the EU and China. Generative artificial intelligence may play an important role in expediting EU–China negotiations. The utilisation of generative AI could contribute to translations of joints statements, communiqués and negotiated texts, allowing fast dissection and diagnosis of each negotiation partner’s intentions, and could even aid in understanding of the latest and trends of foreign direct investment flows between the EU and China.

Secondly, if generative AI raises shared concerns for the EU and China, and if their communications become more efficient, the EU and China may come to reach some kind of shared agreement or framework for generative AI development and governance. This would not be all inclusive and would not be a remedy to all ills but even a minimal agreement would be a real breakthrough. The EU and China are already engaged in a high-level dialogue to facilitate EU–China communication about digital strategy and technological developments: the EU and China established the EU–China High-Level Digital Dialogue in 2020, becoming the fifth pillar of the EU–China comprehensive strategic partnership. The first EU–China High-level Digital Dialogue took place in September 2020, the second EU–China High-Level Digital Dialogue took place in September 2023, where both sides shared updates on their policies and practices in the digital domain, and the EU underscored ‘the importance of an ethical use of this technology in full respect of universal human rights, in the light of recent UN reports. The Commission reiterated its support for global and interoperable Information and Communication Technologies (ICT) standards and urged the Chinese authorities to ensure a fair, reciprocity-based business environment in the digital field’ (European Commission 2023). The Chinese side stressed the opportunities of China’s digital economy development are open for the EU’s engagement to mutual benefits and win–win situation.Footnote 1 If the EU and China could reach a quasi-consensus regarding generative AI governance, through the process of negotiations on AI, and thanks to real time translation, diplomats on both sides could develop trust and a shared language. This could lead to productive talks in other areas and domains such as the trade deficit issue plaguing the EU–China relations or more sensitive cross-strait issues.

Thirdly, in the realm of public diplomacy Generative AI could facilitate the EU and China’s public diplomatic efforts towards each other, in contrast to what many may deem it hindering the relationship. True is the fact that the use of generative AI tools like ChatGPT may augment the spread of disinformation and fake news, especially in authoritarian regimes. But generative AI also presents opportunities to be harnessed in the service of EU–China relationship-building. EU diplomats posted to the EU Delegation to China in Beijing, and especially those in the press section, could utilise generative AI for collecting, arranging and assessing information, for example, to gauge the public perception of the EU’s image and impact in China in a far more efficient manner than the traditional tools adopted by overworked and under resourced public diplomacy practitioners, the responsible use of generative AI systems in public diplomacy could shorten policy loopholes and become a force for good.

Conclusion

The EU and China both play a role in ‘defining how global technological developments will go forward’ (European Commission 2020a). This article has grappled with the question concerning whether AI, and generative AI in particular will escalate tensions between the EU and China or, counter-intuitively reduce tensions? The paper contrasted the two generative AI models adopted by the EU and China, and examined the implications of generative AI for the diplomacy of EU–China relationship-building, explaining how generative AI alters diplomatic outcomes and explored ways of channelling such transformation into warranted directions in diplomatic practice.

Technological capabilities in AI constitute considerable diplomatic power. The pace of generative AI advancements often outstrips regulatory measures; generative AI governance has become a new diplomatic tool, the EU and China should take the lead in defining ‘guardrails’ for generative AI development to create an international forward-looking, risk-based approach to generative AI development and deployment that maximises benefits while mitigating potential risks. AI in general, and generative AI in particular, may be a way to ease tensions globally and to restart the wheels of mutually beneficial relationships as opposed to great power rivalries.