The power of generative AI for CRIS systems: a new paradigm for scientific information management

CRIS systems have evolved significantly over the decades, from basic institutional repositories in the 1970s-80s to incorporating standards, open access, and metrics by the 2000s. Today, CRIS systems aim to store highly interoperable, well-classified data to enable AI integration. Something similar has happened with the evolution of AI systems since the concept was created in 1956.

AI is generally classified into weak AI (narrow, specialized) and strong AI (artificial general intelligence, still theoretical). The EU recently passed the AI Act, the world's first AI regulation, to ensure safe and responsible use while promoting innovation. Spain has its own AI strategy since 2020.

AI will revolutionize research by optimizing funding allocation, accelerating outputs, setting standards, providing infrastructure, addressing legal aspects, and enabling collaboration. Common AI techniques used in CRIS include generative AI for natural language processing, machine learning for text classification, and neural networks for summarization.

AI can automate tasks like researcher profile classification, research groups clustering, SDG alignment, and CERIF semantic enrichment. The IntelComp project demonstrates an AI platform for analyzing large volumes of science, technology and innovation data to assist policymakers, funders and researchers.