FAIR Liquidity Unifying Interoperable Data and AI

The FLUID-AI project introduces a new approach to address the lack of interoperability between data, AI/ML models and solutions within the EOSC. We introduce the concept of Data and Models Liquidity, building on and extending the FAIR principles to address the unique demands of AI-ready data and models. While the FAIR principles have improved data management, they fall short in supporting AI applications, which require data that are not only FAIR but also structured, annotated, and optimized for seamless integration into AI/ML workflows.FLUID-AI identifies and addresses 3 major gaps within the EOSC ecosystem. First, we establish a collaborative Competence Centre (CC) to provide coordinated support, training, and resources, ensuring researchers and operators are equipped with the skills needed to leverage AI/ML tools effectively. Secondly, we promote unified data and models integration, implementing semantical and technical interoperability to enable effortless reuse and combination across platforms and scientific disciplines. Thirdly, we deliver accessible and intuitive platforms, reducing technical complexity so researchers can focus on scientific discover.The project is organized in 3 different action pillars corresponding to the identified gaps. Together with 8 real-world use cases from representative Research Infrastructures would allow us to demonstrate the FLUID-AI impact, validating the project’s solutions, ensuring they are scalable, reproducible, and aligned with real-world research needs.By promoting cross-disciplinary collaboration, standardization, and open science principles, FLUID-AI aims to transform the EOSC into a dynamic, AI-ready ecosystem. The project outcomes include the novel Data and Models Liquidity concept and framework, innovative tools and platforms, comprehensive guidelines, and a blueprint for trustworthy AI-ready repositories. All together will empower researchers to leverage the full potential of AI-driven scientific discovery.