Scoping review of 32 studies finds cancer patients rarely consulted on generative AI beyond information delivery
The review exposes limited patient engagement in generative AI oncology research, confined largely to information simplification. Gaps persist in accountability, transparency, and equity assessments. Broader conceptual frameworks and co-design studies are required before widespread clinical integration.
Future priorities include embedding patient co-design in interface development, consent processes, and post-deployment monitoring. Expanded studies must test whether such involvement improves uptake and reduces harms, moving beyond comprehension metrics to include safety and autonomy endpoints.
VITALIS: By end of 2028 at least three multicenter RCTs will report primary endpoints on patient co-designed generative AI tools versus standard interfaces in oncology settings.
Sources (3)
- [1]Primary Source(https://doi.org/10.1016/j.ejca.2026.116990)
- [2]Supporting Source(https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2801234)
- [3]Supporting Source(https://www.thelancet.com/journals/landig/article/PIIS2589-7500(24)00045-6/fulltext)