Towards Artificial General Intelligence (AGI) for healthcare

Expected Outcome: This topic aims at supporting activities that are enabling or contributing to one or several expected impacts of destination “Developing and using new tools, technologies and digital solutions for a healthy society”. To that end, proposals under this topic should aim to deliver results that are directed at, tailored towards and contributing to all the following expected outcomes: Researchers and innovators benefit from an improved understanding of how to develop and use the next generation of frontier Artificial Intelligence (AI) models for healthcare, including how to leverage AI Factories and how to combine and expand the capabilities of existing foundation models towards inclusive and personalised medicine. Researchers and innovators benefit from an improved understanding of how to leverage highly heterogeneous and multimodal health data spanning a range of anatomical scales (i.e. the micro to the macro level). Multidisciplinary stakeholders have access to a collaboratively created roadmap for developing the next generation of frontier AI models for healthcare, towards Artificial General Intelligence (AGI) for healthcare. Scope: The AI Continent Action Plan [1] identifies the health sector, encompassing life sciences, medical devices and healthcare delivery, as one of the key strategic sectors. The action will contribute to making European life sciences [2] and healthcare more impactful and productive by fostering the full integration of advanced AI in the health sector and biomedical research, along the objectives of the AI in Science strategy [3] and Apply AI strategy [4] . Healthcare typically involves the combining of multimodal data, ranging from electronic health records through imaging and laboratory to molecular and omics data. This information combination is performed by specialists and is often challenging towards optimised patient care. In Europe, the growing amount of accessible multimodal health data, including via the forthcoming European Health Data Space (EHDS) [5] , combined with the increasing availability of high-performance computing facilities (e.g. AI Factories), presents a unique opportunity to develop the next generation of frontier AI models for healthcare. This action anticipates and operationalises the use of such federated infrastructures for research and innovation. Moreover, regulations such as the EHDS regulation and AI Act [6] steer the direction into building an ecosystem fostering ethical and safe innovation on AI in healthcare. AI models are becoming increasingly complex and able to tackle increasin…