From Apertus to the Clinic: Open Foundation Models for Global Health

Switzerland's Apertus is one of the few fully open large language models trained at scale, built on the Alps supercomputer by the Swiss AI Initiative. Imanol Schlag (ETH Zurich) and Martin Jaggi (EPFL) present how Apertus was built and where it is headed. Annie Hartley (LiGHT lab, EPFL) shows what open models enable in practice, building leading open medical LLMs: humanitarian deployments in low-resource hospitals and the MOOVE initiative, which validates medical AI with clinicians worldwide. Together, they make the case that openness is not a constraint but a prerequisite for trustworthy AI in high-stakes domains.