As organizations (and nations) seek greater control over their AI technologies and research processes, the concept of "sovereign scientific AI" has emerged, questionning the possibility of developping robust, independent AI capabilities for scientific discovery while maintaining appropriate control over data, infrastructure and IP in their development strategy.
As AI systems grow more capable in modeling complex biological systems, drug discovery is entering a new era — one where machine learning models don't just assist researchers, but actively generate hypotheses, design molecules, and predict therapeutic outcomes. This session explores the shift from traditional trial-and-error approaches to AI-driven methodologies that can integrate biological data at unprecedented scale and speed. What does it mean to have AI as a true collaborator in science? Can these systems unlock treatments for diseases that have eluded medicine for decades? As AI takes on a greater role in healthcare decision-making, how do we ensure that its use remains transparent, equitable, and ethically sound?