Drawing on scientific literature, strategic documents, white papers and roadmaps, it examines how AI supports the scientific workflow—from literature synthesis and experimental design to data analysis, modelling, discovery, automation, validation and dissemination. It assesses established and emerging AI technology and methods, including machine learning and deep learning, foundation models and generative AI, physics-informed and hybrid models, knowledge graphs and reasoning systems, simulation, optimisation and control, digital twins, and cross-cutting enabling technologies and infrastructures. It also highlights common benefits of AI use in scientific research, including faster analysis and simulation, improved prediction, data integration, more efficient experimentation and increasingly automated workflows, alongside persistent challenges concerning data quality, scientific validation, interpretability, uncertainty, reproducibility, interoperability, computational resources, sustainability, governance and human oversight.
The findings provide an evidence base for identifying research and innovation priorities within SCIANCE and for developing a European Strategic Research and Innovation Agenda for trustworthy, sustainable and competitive AI-enabled science.
Disclaimer: This project deliverable has been submitted to the European Commission and is pending formal review and approval by HaDEA
