Search

Begin New Search Topic Artificial Intelligence Political Geography Europe

1. Which companies are ahead in frontier innovation on critical technologies? Comparing China, the European Union and the United States

Author:
Alicia Garcia-Herrero, Michal Krystyanczuk, and Robin Schindowski
Publication Date:
05-2025
Content Type:
Working Paper
Institution:
Bruegel
Abstract:
Competition in critical technologies is attracting increasing attention not only because of the foundational nature of these technologies for other types of innovation, but also because of their role in the United States national security strategy. In this paper, we look into which entities in China, the European Union and the US innovate at the technology frontier in the three most important critical technologies – artificial intelligence, quantum computing and semiconductors – based on identification of the most radical novel patents in these technologies and their subsectors. Working with these pathbreaking patents, we look into the origin of the companies that file the largest numbers of them. US innovators dominate the innovation frontier for quantum computing and, to a lesser extent, AI, with Chinese innovators doing better in semiconductors. European innovators lag in all, but perform relatively better in quantum computing, in which they rank similarly to Chinese innovators. Furthermore, the innovation ecosystem is quite different across geographies. In the US, tech companies top the rankings of critical novelties and are highly concentrated: as many as three companies are in the top rankings of all of the three critical technologies. Frontier innovators in the field in which the EU competes most equally – quantum – are mostly research centres and not companies. China lies somewhat in between in all three domains.
Topic:
Science and Technology, European Union, Innovation, Artificial Intelligence, Semiconductors, and Quantum Computing
Political Geography:
China, Europe, Asia, and United States of America

2. How DeepSeek has changed artificial intelligence and what it means for Europe

Author:
Bertin Martens
Publication Date:
03-2025
Content Type:
Policy Brief
Institution:
Bruegel
Abstract:
By mid-2024, artificial intelligence large language models (LLMs) were running into diminishing returns to scale in training data and computational capacity. LLM training began to shift away from costly pre-training to cheaper fine-tuning and allowing LLMs to ‘reason’ for longer before replying to questions. Fine-tuning uses chain-of-thought (CoT) training data that includes questions and the logical steps to reach correct answers. This increases the efficiency of learning for smaller AI models, such as DeepSeek. CoT data can be extracted from large ‘teacher’ LLMs to train small ‘student’ models. These changes shift the cost structure of AI models from high pre-training costs to lower fine-tuning costs for model developers and more inference costs for users. While smaller models are cheaper to use, a positive AI demand effect is likely to exceed the negative price effect. Price competition between models will increase, resulting in tighter margins for AI firms. Specialised models can still fetch premium prices. Cheaper LLMs are an opportunity for European Union companies to catch up in building smaller AI models and applications on top of LLMs. Increased demand for AI services will require more investment in computing infrastructure, including in the EU. Investing in large LLMs and the corresponding hyperscale infrastructure is riskier, especially as price competition between models increases. Knowledge extraction between AI models puts pressure on model developers to protect their investments against free-riding by others. It also creates a dilemma for policymakers: should they favour free-riding to promote competition and innovation, or should they clamp down and reinforce monopoly rents to stimulate investment in AI models? Past policy will not be an appropriate response in a world that offers vastly expanded opportunities for knowledge pooling and innovation at lower cost.
Topic:
Science and Technology, European Union, Economic Policy, Microeconomics, Innovation, Artificial Intelligence, Large Language Models (LLMs), and DeepSeek
Political Geography:
Europe

3. Economic arguments in favour of reducing copyright protection for generative AI inputs and outputs

Author:
Bertin Martens
Publication Date:
04-2024
Content Type:
Working Paper
Institution:
Bruegel
Abstract:
Generative artificial intelligence (GenAI) models have stirred considerable controversy about copyright protection for AI training inputs and model outputs. The European Union’s AI Act will require model developers to be transparent about their use of training inputs such as text, images and music. The EU Copyright Directive allows free text and data mining of these media inputs unless copyright holders have opted out and want license payments. The right to opt-out amounts to economically inefficient overprotection of copyright. Free use of media content for GenAI training does not affect media sales to consumers. Optouts only strengthen the bargaining position of copyright holders, who decide depending on their private interests. That generates windfall profits without any increase in consumer surplus or social welfare. The licensing of training inputs reduces the quantity of data and the quality of GenAI models, creates transaction costs and reduces competition between GenAI firms. This slows down GenAI-induced innovation in media products and production processes, and productivity gains in all service sectors that apply GenAI. Ultimately, it slows down economic growth compared to what it could be with competitive and high-quality GenAI. Bargaining over license pricing is arbitrary as there is no objective revenue benchmark to start from. Defenders of the moral right to remuneration argue that any arbitrary remuneration is better than no remuneration. But this private moral right comes at the expense of social welfare. The ongoing bargaining and court cases between media producers and GenAI developers risk entrenching this market failure in jurisprudence. Early regulatory intervention and elimination of opt-outs for GenAI training, or weakening them in AI Act implementation guidelines, would solve this. There is no need for copyright on GenAI outputs either. GenAI reduces the marginal cost of machine-production of media outputs to close to zero, on par with the marginal cost of reproduction. That eliminates incentives for piracy. Moreover, composite human-machine outputs benefit from a de-facto extension of copyright on the human component.
Topic:
Science and Technology, Intellectual Property/Copyright, European Union, Regulation, Economic Growth, and Artificial Intelligence
Political Geography:
Europe

4. Exposure to generative artificial intelligence in the European labour market

Author:
Laura Nurski and Nina Ruer
Publication Date:
03-2024
Content Type:
Working Paper
Institution:
Bruegel
Abstract:
We apply two sets of generative artificial intelligence (GenAI) occupational exposure scores – one task-based, one ability-based – to the European Labour Force Survey. While using different methodologies, our findings reveal consistent demographic patterns across the two approaches: jobs held by women, highly educated and younger workers are more exposed to GenAI technology in Europe. We also review the literature on the recent productivity impact of GenAI. Within the same occupations, less-experienced or less-skilled workers consistently get the largest productivity gains from GenAI support. We argue that a task-based analysis is more fruitful than an ability-based one, both for guiding GenAI adoption in organisations and their workplaces, and for assessing the employment and job quality impact on workers. Finally, we provide policy recommendations that can help workers (ie the labour supply) adjust to technological disruption, such as providing training and social safety nets. But we go further by also suggesting policy interventions that could redirect future labour demand towards better jobs, by promoting job redesign and organisational agility. Monitoring GenAI’s employment effects and researching the ‘jagged technological frontier’ is necessary to further build our understanding of the employment impact of this transformational technology.
Topic:
European Union, Artificial Intelligence, Labor Market, and Supply and Demand
Political Geography:
Europe

5. Exploring the Application of AI in the Public Sector: The Case of Estonia and Lessons for Georgia

Author:
Salome Abramishvili
Publication Date:
08-2024
Content Type:
Commentary and Analysis
Institution:
Georgian Foundation for Strategic International Studies -GFSIS
Abstract:
Artificial Intelligence (AI) is rapidly transforming societies globally, offering unprecedented opportunities across sectors. The integration of AI into the public sector promises enhanced service delivery, efficiency, and significant decision-making improvements. Nations worldwide recognize AI’s strategic importance and are adopting National AI Strategies (NASs) to guide its development and deployment. This paper explores AI’s evolution and impact in the public sector, focusing on applications and strategic frameworks using Estonia as an example—a pioneer in AI governance within the EU. Estonia’s approach serves as a model for Georgia, a country with post-Soviet roots and EU integration aspirations. By examining Estonia’s National AI Strategy and its implementation, this paper identifies key lessons and recommendations for Georgia to develop its tailored strategy.
Topic:
Science and Technology, Public Sector, and Artificial Intelligence
Political Geography:
Europe, Estonia, and Georgia

6. Nuclear arms control policies and safety in artificial intelligence: Transferable lessons or false equivalence?

Author:
Eoin Micheál McNamara
Publication Date:
01-2024
Content Type:
Policy Brief
Institution:
Finnish Institute of International Affairs (FIIA)
Abstract:
Multiple nuclear arms control treaties have collapsed in recent years, but analogies associated with them have returned as possible inspiration to manage risks stemming from artificial intelligence (AI) advancement. Some welcome nuclear arms control analogies as an important aid to understanding strategic competition in AI, while others see them as an irrelevant distraction, weakening the focus on new frameworks to manage AI’s unique and unprecedented aspects. The focus of this debate is sometimes too narrow or overly selective when a wider examination of arms control geopolitics can identify both irrelevant and valuable parallels to assist global security governance for AI. Great power leaders frequently equate AI advancement with arms racing, reasoning that powers lagging behind will soon see their great power status weakened. This logic serves to intensify competition, risking a spiral into more unsafe AI practices. The global norm institutionalization that established nuclear taboos can also stigmatize unethical AI practices. Emphasizing reciprocal risk reduction offers pragmatic starting points for great power management of AI safety. This research is part of the Reignite Multilateralism via Technology (REMIT) project, funded by the European Union’s Horizon Europe research and innovation programme under Grant Agreement No. 101094228.
Topic:
Defense Policy, Arms Control and Proliferation, Cybersecurity, Economic Policy, Artificial Intelligence, and Strategic Competition
Political Geography:
Russia, Europe, and United States of America

7. The EU and military AI governance: Forging value-based coalitions in an age of strategic competition

Author:
Katja Creutz, Ville Sinkkonen, Mahmoud Javadi, and Michal Onderco
Publication Date:
02-2024
Content Type:
Working Paper
Institution:
Finnish Institute of International Affairs (FIIA)
Abstract:
The use of Artificial Intelligence (AI) has increased dramatically, driving discussion on the best way to govern it. While much of the discussion has focused on AI in the civilian domain, its usage in military applications raises distinct ethical and legal questions that actors increasingly seek to address. Governance initiatives are becoming more common, even though strategic competition between major powers remains a challenge. The EU navigates this governance landscape more hesitantly than in the civilian domain with its Artificial Intelligence Act (AIA). More information is needed regarding what kind of regulation, and especially with whom the Union could pursue such avenues. This paper seeks to address these questions by presenting and analyzing results from an expert survey conducted in autumn 2023 by researchers at Erasmus University Rotterdam. It argues that the EU needs to pursue a broad range of cooperation formats to advance its values and interests in the context of military AI. Strategic multilateralism is needed, especially with like-minded states that share values with the Union.
Topic:
Science and Technology, Governance, European Union, Artificial Intelligence, and Strategic Competition
Political Geography:
Europe

8. AI Unleashed or Tamed? Decoding Europe’s Bold Leap with the AI Act

Author:
Arianna Jater
Publication Date:
03-2024
Content Type:
Policy Brief
Institution:
International Centre for Defence and Security - ICDS
Abstract:
2023 saw a boom in generative Artificial Intelligence (AI). However, such a rapid expansion revealed the unpreparedness of governments to handle the unprecedented impact AI systems have on the core values of democratic societies. The European Commission’s AI Act has emerged as the first binding legislation of its kind. It is already clear, however, that simply ratifying the Act will not solve the next challenges. Among those, the immediate one is to ensure that the new AI rulebook is enforced on the national level and that Member States have the resources necessary. Then, the Act, and any future AI legislation, will need to constantly adapt to the ever-changing developments in technology, which now seem to move at a faster pace than ever before. Interdisciplinary approaches and new research disciplines will be needed to analyse the correctness and safety of AI systems, as well as to evaluate their impact on society. Last but not least, the private sector — already employing or only exploring the possibility of introducing AI systems — will need strong support to remain compliant. This brief outlines the state of the EU’s AI Act, discusses concerns within Europe, examines the position taken by the Estonian technology sector and government, and explores the future direction the regulation might take.
Topic:
Science and Technology, European Union, Regulation, Legislation, Artificial Intelligence, Private Sector, and European Commission
Political Geography:
Europe and Estonia

9. Russia’s War in Ukraine: Artificial Intelligence in Defence of Ukraine

Author:
Vitaliy Goncharuk
Publication Date:
09-2024
Content Type:
Policy Brief
Institution:
International Centre for Defence and Security - ICDS
Abstract:
Russia’s full-scale war in Ukraine is the first international conflict in which the opposing sides have actively developed and used artificial intelligence (AI) for military purposes. AI solutions for geospatial intelligence, operations with unmanned systems, military training, and cyber warfare have been key to success on the battlefield. This brief explores the state of defence AI in Ukraine and highlights the main lessons that can be learned from the Ukrainian experience.
Topic:
Defense Policy, Science and Technology, Artificial Intelligence, and Russia-Ukraine War
Political Geography:
Europe and Ukraine

10. AI Policy in EU Illiberal Democracies: The Experience in Hungary and Poland

Author:
Filip Konopczyński
Publication Date:
01-2024
Content Type:
Special Report
Institution:
German Marshall Fund of the United States (GMFUS)
Abstract:
This paper examines the emergence of AI policies in Hungary and Poland under illiberal governments, and highlights their potential social and political consequences, particularly for democratic values and civil and fundamental rights. It focuses on the adoption of AI in the public sector, encompassing research and development, public administration, law enforcement, migration, and economic policy. In their AI policies, both countries’ governments have prioritized industry demands and subordination to the expectations of large foreign corporations (which is inconsistent with their digital sovereignty rhetoric). Meanwhile, they have neglected societal consultations and the needs of the scientific community. The AI policies implemented in Hungary and Poland by the Fidesz and Law and Justice (PiS) parties) have been characterized, respectively, by centralization and fragmentation, with varying outcomes. The AI systems deployed do not safeguard citizens’ rights as the political takeover of the justice system and partisan control of law enforcement have undermined redress mechanisms and limited legal protection from AI-related violations. The increasing use of AI in election campaigns, coupled with the lack of democratic oversight, increases the risk of mass disinformation campaigns and electoral manipulation in both countries. The cases of Hungary and Poland highlight some key implications for democracy and human rights in the EU where illiberal actors control AI policies and governance and disregard these values. The new EU AI Act may offer some protection for the rule of law and individual rights, but its potential loopholes could allow the unlawful deployment of AI systems in vital areas. AI policies in both countries have reflected their governments’ illiberal tendencies, expanding their control over citizens and curtailing democratic processes. The centralized governance raises concerns about the potential for mass surveillance and censorship, while the lack of transparency and inclusivity in AI policymaking could further marginalize minority groups and vulnerable populations. However, there are key steps that the EU can take to address these issues.
Topic:
Science and Technology, European Union, Domestic Politics, Artificial Intelligence, and Illiberal Democracy
Political Geography:
Europe, Poland, and Hungary