Looking ahead

Europe can lead a new wave of progress in AI technologies, creating next-generation AI that is made in Europe and that delivers real-world benefits for communities and businesses. To deliver this European capability, Europe needs to be at the forefront of advancing AI technologies, deploying them in service of European priorities, and connecting their development to the rights and values set out in European law. In pursuing these goals, Europe needs to grow its AI R&D ecosystem, creating a pipeline from development to deployment through collaborations that embed knowledge creation in local innovation ecosystems.

The transformative potential of AI stems from its pervasiveness. AI has been posited as the next General-Purpose Technology, with the ability to disrupt economies and societies, bringing both benefits and risks. The history of General-Purpose Technologies shows that their effects are felt over decades, through waves of technological progress and shifting patterns of adoption. Over the coming decades, the challenge for policymakers is to adapt to these shifting patterns; to steward a technology that is changing, pervasive across sectors, and enmeshed with societal interests and concerns. To position itself for rapid response, Europe can act to strengthen its AI foundations, by investing in research that advances core underlying technologies while advancing the trustworthiness of AI in deployment, and by building a sustainable infrastructure for AI research.

Recent advances in Large Language Models signal the transformational change that rapid progress in AI technologies could bring. Today’s applications of these foundation models scratch the surface of AI’s potential: the analysis of the use of AI for the EU’s Innovation Missions presented in this report showcases what AI could help achieve across priority policy areas. Progress in foundation models also demonstrates the importance of strengthening the fundamental infrastructure underpinning Europe’s AI ecosystem.

This progress has been driven by innovations in core underlying technologies, and investment to secure European leadership in these technologies is vital. Ensuring such progress delivers benefits for people and society requires human-centric AI methods and applications, through research, policy, and practice that connects technology development to the rights set out in European law.

Translating these technologies to wider economic and social benefit relies on an ecosystem that facilitates AI adoption, through wider access to AI skills and a start-up and business environment that can deploy AI safely and effectively. Support to continue to build this infrastructure, and embed these functions in local innovation ecosystems, is necessary to position Europe to capitalise on recent AI advances and lead technology development in the future.

As a Network of Excellence, ELISE has shown how the EU’s investments in AI can advance a strategically important research agenda, attract leading research talent to Europe, and translate research into innovation. The result is both enhanced collaboration across Europe and local connections that ensure communities and businesses benefit from AI progress. Sustained investment can ensure this ecosystem continues to grow, delivering AI made in Europe, for Europe and the wider world.

ELISE Research Programmes and Use Cases

ELISE currently convenes the following research programmes:

Machine Learning for Health
Geometric Deep Learning
Interactive Learning and
Interventional Representations
Machine Learning for Earth
and Climate Sciences
Natural Intelligence
Quantum and Physics-Based
Machine Learning
Semantic, Symbolic and
Interpretable Machine Learning
Robot Learning:
Closing the Reality Gap!
Human-centric Machine Learning
Machine Learning and
Computer Vision
Multimodal Learning Systems
Natural Language Processing
Robust Machine Learning
Theory, Algorithms and Computations
of Modern Learning Systems

Enhancing each of these work programmes, ELISE facilitates a suite of industrial collaborations around the following use cases:

Environment Perception for Autonomous Driving
AI Explainability for Optical Inspection in Manufacturing
Environment Perception for Autonomous Driving
Generative Adversarial Networks for Real-time Rendering
Robust and Certifiable Multi-Modal Learning for Safe Human-Robot Interaction
Data-Efficient Activity Recognition in Video
Audio Representations in Hearing Health Care
Algorithmic Validation of Smart City AI System Behaviour
Knowledge Scene Graphs for Industrial Applications
Material Flow Optimization
Experimental Environment for Real World Reinforcement Learning

Participants in Innovation Mission workshops

Thank you to the researchers, policymakers, and practitioners that contributed to ELISE’s workshops on AI and the EU Innovation Missions in Winter 2022 and Spring 2023.

View participant list

Acknowledgements

In producing this refresh of ELISE’s Strategic Research Agenda, we are grateful for contributions made by:

  • Leaders of ELISE research programmes who contributed to surveys and workshops
  • Members of the research and policy communities who contributed to workshops on AI in the Innovation Missions
  • The Oxford Insights team that supported those workshops through a literature review and stakeholder interviews on AI and the Innovation Missions, and the individuals who agreed to be interviewed in this process
  • The Collective Next team that facilitated workshops
  • Reviewers from the ELISE and ELLIS networks who helped develop its content

European Learning and Intelligent Systems Excellence (ELISE) is a European Network of AI Excellence Centres funded by the European Commission under the Horizon 2020 Framework Programme under Grant Agreement no 951847. ELISE began in September 2020. The information contained in this document reflects only the author’s views and the Community is not liable for any use that may be made of the information contained therein. For more information on ELISE, please see www.elise-ai.eu.

Copyright notice: Copyright © Members of the ELISE Consortium, 2023.

This document was produced with support from: Sami Heinäsmäki (Aalto University); Samuel Kaski (Aalto University and University of Manchester); Neil Lawrence (University of Cambridge); Jessica Montgomery (University of Cambridge); Christopher Murray (Aalto University); Sue Sebborn (University of Cambridge); and Arnold Smeulders (University of Amsterdam).

Back to top
Back to start