The next era of medicine will depend not only on more powerful AI, but on whether the data, evidence, and knowledge beneath it can be made computable, connected, trustworthy, and continuously improving — so that AI advances care effectively, ethically, and for the benefit of all.
Biomedical knowledge should be represented in forms that people and machines can use, and linked across discovery, care, public health, and education — not locked away in disconnected publications, records, and systems.
AI is only as trustworthy as the data, evidence, assumptions, and governance beneath it. Trust has to be engineered into both the knowledge and the models — and verified continuously, not assumed.
Real-world data should constantly produce real-world evidence — so that what we know keeps pace with how medicine is actually practiced, and discovery and translation accelerate.
Biomedical knowledge and AI matter only insofar as they improve lives — translating into better care, faster discovery, and stronger health, delivered effectively, ethically, and equitably for individuals and populations alike.
Care should generate evidence, evidence should improve care, and — increasingly with AI — that cycle should become routine, turning every encounter into a chance to learn.