This session offers a platform to discuss current developments in the field of data-driven methods and machine learning, which are transforming material modeling and computational mechanics more broadly. We invite contributions including but not limited to the development of machine learning-based surrogate and large-scale foundation models, as well as model-free methods that directly integrate material data to accelerate mechanical, multiscale and multiphysics simulations. Particular emphasis is placed on the development of methods that embed physics and thermodynamics priors within data-driven techniques, enhancing their robustness and reliability. Techniques of interest also involve interpretable/explainable and multifidelity methods that reduce the need for large amounts of training data, as well as probabilistic and uncertainty quantification approaches. We finally invite researchers to contribute their findings in application-focused studies in solid mechanics, biomechanics, geomechanics and related disciplines. The minisymposium is meant as an interdisciplinary platform to discuss the future of data-driven modeling and its potential in complementing traditional approaches in mechanics.