Data-Centric AI Manifesto: How Data Quality Drives Modern AI

This white paper represents the joint outcome of five partners within the FAIR project, spanning four different spokes, who collectively contributed to the activities of Transversal Project 7 (FAIR TP7 – Data-Centric AI and Infrastructures). The contributing institutions are: University of Bari Aldo Moro (Spoke 6), University of Calabria (Spoke 9), University of Naples Federico II (Spoke 3), University Sapienza of Rome (Spoke 5), and INFN – National Institute of Nuclear Physics (Spoke 6). Their coordinated effort reflects the interdisciplinary and multi-institutional nature of FAIR, reinforcing the central role of data-centric methodologies in advancing trustworthy and robust AI systems.