Cancer Research UK funds three teams to test AI co-scientists on major unsolved cancer questions
Three research teams are using large language model-based AI tools as collaborators to generate new experimental hypotheses from existing scientific literature and datasets.
Cancer Research UK has published details of three funded teams exploring whether AI tools — described as "co-scientists" — can identify research directions and experimental hypotheses that might otherwise be missed by human researchers working through the scientific literature and existing datasets.
The initiative reflects growing interest across the research community in using large language model-based systems to synthesise published evidence at scale, surface cross-domain connections, and suggest candidate experiments. The specific cancer questions being addressed by the three teams were not detailed in the feed lede available to Genetic Current.
The term "co-scientist" has entered research parlance through commercial and academic programmes, most prominently a 2025 publication from Google DeepMind describing an AI system designed to assist with experimental reasoning in biology. Cancer Research UK's programme positions the AI tools as aids to researcher thinking, not autonomous agents, consistent with current norms in responsible AI research deployment.
The initiative does not involve AI being used to interpret patient data or inform treatment decisions; it is framed as a tool for scientific hypothesis generation. How well such systems perform — and whether the resulting hypotheses translate into productive experiments — remains an open empirical question. Cancer Research UK is the primary funder of this initiative and published the news on 6 November 2026.
Sources
Read the original reporting — these are the public sources this summary draws from.
-
Primary source Cancer Research UK · 2026-11-06Can AI "co-scientists" help answer cancer's biggest questions?