Council calls for an inclusive, ethical, sustainable and human-centric strategy for the uptake of AI

The Council has approved a set of conclusions titled ‘Towards the EU strategy on AI in science’. In this document, the Council calls for a comprehensive strategy to boost the uptake of AI in science, with an ethical, sustainable, inclusive and human-centric approach.

Artificial intelligence can completely transform scientific discovery, boost innovation, and enhance the productivity of our researchers. Europe must seize these opportunities quickly, while also addressing the ethical and societal challenges and risks of this rapidly advancing technology.

Marcin Kulasek, Polish minister of Science and Higher Education

A game change technology

The Council conclusions recognise that the rapid development of AI and its applications in scientific research has led to groundbreaking achievements, transforming the way science is practiced. AI is used extensively in data analysis, simulations, and large language models, driving new discoveries and enhancing research efficiency. The conclusions underline the importance of AI in supporting both basic and applied research, and its critical role in addressing global challenges and improving competitiveness, societal needs, and the digital transition in Europe.

AI collaboration, especially reciprocal and non-discriminatory partnerships at the international level, is key to enhancing scientific exchange, ensuring interoperability, and promoting responsible AI development. The conclusions emphasise the importance of ethical, sustainable, and inclusive AI use in science, which will help strengthen research, innovation (R&I), and global competitiveness. The text recalls that AI systems and models specifically developed and put into service for the sole purpose of scientific research and development are exempt from the AI Act.

Time to act

Despite the growing significance of AI in science, the EU currently lacks a dedicated policy to facilitate the uptake of AI in science, the conclusions warn. A new European strategy for AI in science is under development by the Commission, which the conclusions emphasise should be based on the best available knowledge and practice and developed in close cooperation with the member states and the R&I community. In particular, the strategy should complement existing initiatives and focus on interdisciplinary research ecosystems, coordinated policy developments, upskilling researchers, and ensuring open access to reliable data. This strategy should also emphasise the responsible and ethical use of AI tools in research and innovation (R&I), ensuring privacy, data integrity, and security. Furthermore, the text calls for a common European agenda to support AI research communities and improve access to data, computational resources, and scientific talent. The conclusions take note of the Commission’s idea to establish a European AI Research Council, while calling the Commission to work with member states on the details of this initiative, in particular its mission and governance to make the best use of existing initiatives and structures.

AI skills for the scientific community

The conclusions also stress the importance of upskilling and reskilling in AI for researchers and innovators, aiming to bridge the digital skills gap through education and training programmes. This includes initiatives to attract, retain, and bring back AI talent to Europe, especially from underrepresented groups, and foster diversity and gender balance in STEM (science, technology, engineering and mathematics) fields.

Risk control

AI adoption in science also presents risks such as biased algorithms, data manipulation, and ethical concerns, which could lead to misinformation or disruptions. To mitigate these risks, the Council urges the development of quality benchmarks, guidelines, and best practices for AI use in research, ensuring integrity, transparency, and reproducibility of scientific work. The Council encourages open data sharing in line with FAIR (findability, accessibility, interoperability, and reusability) principles, promoting data accessibility, interoperability, and ethical AI practices.

In conclusion, the Council calls for enhancing the integration of AI in science through policy support, fostering collaboration, ensuring ethical standards, and providing resources for AI training.

Background

According to the Commission’s Group of chief scientific advisors, AI is one of the most disruptive sets of technologies and is quickly becoming a transformative tool for scientific processes. The share of scientists using AI in research increased from 12% in 2020 to 16% in 2021. In the European Research Council, the use of AI covers almost all fields. In 2021-2022, about one quarter of AI projects in the Joint Research Centre used AI related technologies.