Google DeepMind's Co-Scientist system is being applied to liver disease research, with two new case studies demonstrating how the AI can generate plausible hypotheses for complex medical problems. In one study, researcher Filippo Menolascina used the system to investigate why certain existing drugs help only a subset of patients, aiming to uncover the underlying disease mechanisms that drive those differences.
The second study, led by a Stanford geneticist, focused on finding repurposed medicines for chronic liver disease and liver fibrosis. Instead of searching for entirely new molecules, the AI helped identify existing drugs that might be reused against fibrosis, a common and serious consequence of progressive liver damage.
Both studies share a common thread: Co-Scientist is being used to speed up the early, hypothesis-heavy stage of therapeutic discovery. They differ in emphasis—one looks at patient variability and drug response, while the other targets repurposing for fibrosis. Together, they suggest the AI tool is flexible enough to assist with both mechanistic questions and drug candidate identification, though the actual clinical benefits remain to be tested.