AI’s growing role in scientific discovery comes with rising computational and energy demands. This talk highlights the need for infrastructure, workforce development, and knowledge integration to ensure AI remains sustainable, explainable, and impactful. We explore AI’s role in climate and health predictions, the energy challenges of large-scale AI, and solutions like efficient workflows and provenance tracking for transparency and reproducibility. Using XFEL experiments as a case study, we demonstrate AI’s potential while advocating for hybrid HPC-Cloud solutions and stronger AI education. This talk discusses the critical need to balance efficiency with trust, maximizing AI’s impact while addressing sustainability challenges.
