The European Artificial Intelligence Act, which came into force in August 2024, aims to foster responsible AI development across the EU. European Commission President Ursula von der Leyen has proposed the establishment of a European AI Research Council, akin to CERN, to bolster AI research and development.

Despite leading in AI regulation, the EU has struggled to cultivate a thriving domestic AI industry, hindered by limited computational resources and funding. Stakeholders argue that the next step involves focusing on trustworthy advanced AI (AAI), which encompasses lawful, ethical, and robust general-purpose models.

At a Euractiv panel discussion, experts deliberated on the concept of a “CERN for AI,” exploring its potential political, geopolitical, and economic impacts.

The panel included Maria Cristina Russo, Director of Prosperity at the European Commission’s DG for Innovation Research; Mika Altola, Member of the European Parliament; Thomas Jorgensen, Director of Policy Coordination and Foresight at the European University Association; Sabine de Mey, Director of Flanders Artificial Intelligence Research Program; and Max Riddell, Advanced AI Director at the Center for Future Generations.

The idea of a centralized AI research hub, akin to CERN, was discussed. While some panellists supported the metaphor, others suggested a more distributed approach to leverage Europe’s strengths in cooperation and innovation.

Cristina Russo emphasized the need for a strategic agenda to reinforce European strengths and competitiveness in AI. She highlighted the importance of listening to stakeholders, including scientists and industry leaders, to develop a robust AI framework.

Trustworthy AI

The panel underscored the importance of developing trustworthy AI, focusing on transparency, explainability, and reliability. Max Riddell and Sabine de Mey highlighted the need for AI systems that are robust and reliable to gain public and industrial trust.

Mika Altola, MEP, pointed out the geopolitical dimensions of AI development, noting that Europe must navigate the competitive landscape dominated by the US and China. He stressed the need for Europe to adopt a unique approach, avoiding mere replication of American models.

The discussion also touched on the broader implications of AI for the economy and society. Sabine de Mey emphasized the need for AI to address societal challenges, including environmental sustainability while ensuring economic viability.

The panellists agreed on the necessity of a balanced regulatory approach that fosters innovation without stifling it. Cristina Russo and Mika Altola advocated for a regulatory framework that supports market openness and provides incentives for AI development.

Trust issues

When it comes to trust and guidelines, the European Commission has said it aims to build citizens’ confidence in AI by promoting guidelines that ensure transparency, explainability, and accountability. These guidelines are intended to be followed by national authorities, universities, and research institutions.

Cristina Russo noted that there is strong political momentum in Europe to act on AI, both in science and general applications. The upcoming Multiannual Financial Framework (MFF) and Horizon Europe program are expected to focus on AI-related objectives.

The discussion highlighted the need for substantial funding to compete with initiatives like the US Stargate project and China’s investments in AI. Public-private partnerships and leveraging seed money were identified as key strategies to multiply resources and achieve the necessary scale.

When it comes to strategic autonomy, MEP Axel Voss emphasized the need for Europe to play a significant role in AI to ensure digital survival and competitiveness. He called for a transformation fund and strategic projects to activate available resources and attract private investment.

The panel discussed the benefits of a dispersed model for AI research and development, leveraging Europe’s strengths in innovation and collaboration. This approach involves coordinating efforts across various institutions and regions, rather than centralizing everything in one location.

They acknowledged the competitive landscape, with the US and China making significant strides in AI. The importance of Europe investing in AI to avoid being left behind was underscored, with a focus on developing a coordinated strategy that includes both foundational technologies and applications.

Aside from the threat of competition, Thomas Jorgensen downplayed the existential threat of AI, suggesting that concerns are often exaggerated. He emphasized the importance of focusing on practical applications and avoiding science fiction scenarios.

Jorgensen also highlighted the importance of applying AI in scientific research, advocating for a distributed model that leverages Europe’s collaborative strengths. He stressed the need to develop a full European AI stack while focusing on practical applications.

Encouraging private investment

Sabine de Mey underscored the necessity of translating research into production to attract private investment. She cautioned against tunnel vision, advocating for the exploration of new AI paradigms beyond regenerative AI and leveraging Europe’s domain knowledge.

Max Riddell pointed out the lack of leading AI companies focused on creating trustworthy systems. He argued that the EU could fill this gap with significant investment, unlocking breakthroughs in science, the economy, and general well-being.

Axel Voss called for a comprehensive strategy to address AI infrastructure needs. He emphasized the importance of viewing AI as a fundamental element for the future and the necessity of substantial financial resources to support this vision.

The panellists spoke of the need for a coordinated approach, leveraging public-private partnerships and focusing on both foundational technologies and applications to ensure Europe’s leadership in AI.

The discussion concluded with a call for a creative and fit-for-purpose approach to AI development in Europe, leveraging the continent’s unique strengths and ensuring that AI serves both scientific and societal goals.