The AI dual-use dilemma using the example of China
This article discusses the dual-use dilemma of AI, focusing on China’s approach and the challenges of balancing innovation with security risks, particularly the blurred lines between civilian and military applications.
Disclaimer: this is an opinion piece. The views and opinions expressed in it are those of the author and do not necessarily reflect the official policy or position of Allegro.
Introduction #
The dual-use dilemma of AI applies to virtually every country that is either already a superpower or aspires to be one. This article could, in principle, describe any major player on the international stage. So why China? Firstly, because its actions are highly visible. Secondly, though perhaps belatedly, I will try to capitalize on the boom associated with DeepSeek – a tool which is probably the most apparent manifestation of Chinese ambitions in AI. Reports of leaked personal databases, chat histories, and other security vulnerabilities in the DeepSeek application may just as well be only the tip of the iceberg.
The Changing Cybersecurity Landscape #
At the risk of being Captain Obvious, I will start this discussion by stating that advances in AI are changing the cyber security landscape dramatically.
The ambitious plans of the Chinese, combined with their military-civilian fusion strategy, present an explosive mix and a challenge that raises concerns about the potential misuse of AI technologies, to say the least. This dual-use dilemma refers to the inherent challenge of balancing the desire to foster innovation and economic growth with the need to mitigate potential security risks. At the end of January, I wrote a few words on Substack about how the European Union is putting a stick in its spokes by introducing the AI Act in its current form and at such a moment in history. Regulation is needed, of course. But moral superiority may not be enough in this day and age.
Many advances in Artificial Intelligence are based on open source research, international collaboration and the free flow of information. This openness fosters innovation and accelerates progress, but it also increases the risk of these technologies falling into the wrong hands. Controlling the dissemination of knowledge and technology while preserving the benefits of international cooperation is a major challenge.
Challenges of Control and Regulation #
Imposing strict export controls on AI technologies, in turn, stifles innovation and economic growth. Such restrictions damage international relations and provoke retaliation from other countries. Moreover, to indulge in another truism, it is tough to effectively control the flow of information and technology in the digital age.
The boundary between civilian, political or even military applications of Artificial Intelligence has always been blurred. Facial recognition technology, for example, was initially developed for convenient payment systems and smart city initiatives, but can be (and is) easily adapted for mass surveillance or tracking of individuals of interest. Similarly, AI-based cybersecurity tools designed to protect critical infrastructure can also be (and are) used to develop sophisticated cyber-weapons and conduct cyber espionage.
Blurred Lines: Civilian vs. Military Applications #
China’s approach makes it difficult to track technology flows and identify potential military applications. Universities and research laboratories may be involved in both civilian AI projects and those with potential military applications, making it difficult to distinguish between the two.
Widely used in China for convenient payments and smart city initiatives, facial recognition technology can also be used for mass surveillance, people tracking and even in autonomous weapons systems. The military could use facial recognition to identify enemy combatants, high-value targets, and even civilians in real-time during conflict. In the future, autonomous weapons systems could use facial recognition to select and attack targets. Analysis of huge datasets of facial images could be used to identify individuals, track their movements, and create detailed profiles for intelligence purposes. China’s social credit system, which relies heavily on facial recognition, can de facto be a system of mass surveillance and social control. This system could easily be adapted for military purposes.
Examples of Dual-Use AI in China #
While AI can be used to strengthen cybersecurity defences, it also poses new challenges, such as the “black box” problem of understanding how AI systems make decisions and identify potential vulnerabilities.
Natural Language Processing (NLP), in turn, can power translation services and customer service, but it can also be used to generate deepfakes, spread disinformation, and analyze vast amounts of textual data to gather intelligence. It can also be used to analyze vast amounts of textual data, such as social media posts, news articles, and enemy communications, to extract valuable intelligence. Artificial Intelligence can be used to analyze public sentiment, identify potential threats, and predict likely population reactions to political events or military actions. NLP can be used to generate and disseminate disinformation campaigns, manipulate public opinion, and sow discord among enemy populations. AI-based tools are already widely used to create highly realistic deepfakes, including manipulated videos and audio recordings of political leaders, which can be used for propaganda, blackmail, and sowing confusion and distrust.
Data Collection and Its Implications #
Artificial Intelligence models are trained on massive datasets. China has a vast amount of data on its citizens, collected through various means, including social media, surveillance systems, and government records. Now, through the release of free apps and open-source tools, they also have data on citizens of other countries. This data advantage can be used for military purposes, such as predictive modeling and the development of superior Artificial Intelligence systems. The collection and use of massive amounts of data raises serious privacy and data security concerns. The Chinese government could misuse this data for social control, surveillance, and suppression of dissent. The potential data breaches and misuse of personal data collected by Chinese companies should be of concern to people around the world. I write “should” because the percentage of people who are concerned about what happens to their data is still too small for us to talk about a breakthrough in the perception of this problem.
The AI Arms Race #
China’s rapid advances in Artificial Intelligence are fuelling an arms race between major powers. This competition is leading to cyberattacks, the development of autonomous weapons systems, and increased geopolitical instability. The rapid development and deployment of AI-based weapons systems may increase the risk of miscalculation and unintended consequences. There will be no single individual to blame. It will simply be decision-makers saying, “Oops, accident, the computer made a mistake, but it’s not our fault.” The development of AI-based technologies could fundamentally change the balance of power on the international stage. China’s dominance in Artificial Intelligence could give it a significant economic and technological advantage, potentially affecting global trade and economic order. The development of advanced weapon systems based on Artificial Intelligence could alter the balance of military power, potentially destabilizing the existing international order. Which, as I wrote earlier, is what China is aiming for.
Finding a Solution #
I am struggling to find a solution to this situation. No wonder: smarter people have tried and failed. Introducing more and more new regulations will not stop the superpowers from developing AI tools in the way that is forbidden in the EU. In my opinion it will be quite the opposite – it will further tie the hands of entrepreneurs and inventors in the EU and cause a further exodus to countries with a more liberal approach to technology. International players who have so far failed to comply with EU regulations will continue to do so.
The only solution I see is raising public awareness of the potential risks and benefits of Artificial Intelligence technologies. Fostering public discourse and debate on the ethical, social, and political implications of Artificial Intelligence development is essential. It is also crucial to invest in education and training programs to equip future generations with the knowledge and skills necessary to navigate an AI-based future.