FAIR&AI: AI and ML potential in RDM: Opportunities and action areas

This is a preprint version of the paper to be submitted. The Research Data Lifecycle spans from research planning to data collection, processing, publication, archiving, and reuse. Each phase presents specific challenges that can be addressed through Artificial Intelligence (AI) and Machine Learning (ML). Based on a systematic mapping of stakeholder involvement, we identify and prioritize AI/ML applications with the highest potential impact on efficiency, quality, and FAIR compliance. We identify strategic action areas where AI/ML could support future shared RDM services and where no mature solutions currently exist. The increasing availability of generative AI and machine learning (AI/ML) methods raises new opportunities and challenges for research data management (RDM) . The Shared RDM Services and Infrastructure project was launched to establish a framework that provides selected tools and infrastructures as shared services for Austrian universities and research institutions. Building on the project’s original focus on interoperability, FAIR principles , and the bundling of expertise across institutions, a dedicated working group was established in 2025 to explore potential intersections between AI technologies and the research data cycle. In this paper, we present a structured mapping of AI/ML opportunities along the stages of the research data lifecycle, complemented by cross-cutting issues, such as metadata interoperability, legal compliance, and quality assurance. For each stage, we identify typical activities, relevant stakeholders, gaps in current practices, potential AI/ML applications, and their expected impact. We highlight “action areas,” such as the absence of mature tools for cross-domain metadata annotation and AI-supported license advising, where new shared RDM services could be developed. AI-based models and tools could serve as an incentive for existing research in the field of AI/ML, while at the same time paving the way for new, innovative approaches to be investigated in a specific sub-field of computational science or AI research. The analysis provides an avenue for both, use cases that lead directly to quick wins and long-term strategies for integrating AI into national and European research data infrastructures. We argue that AI can act as an enabling layer for FAIR and Open Science, if service design, governance, and responsible use remain in focus. The reference list for the literature review is available as a supplementary file in the dataset: 20260415-wos-savedrecs.xls