I found this interesting for my research. A new AI model from Harvard Medical School researchers is showing promise in predicting which cancer patients may respond to immunotherapy. The research also raises a bigger question for healthcare AI: when a model moves closer to influencing patient care, what evidence should we require before we are prepared to act onA new AI model from Harvard Medical School researchers is showing promise in predicting which cancer patients may respond to immunotherapy. The research also raises a bigger question for healthcare AI: when a model moves closer to influencing patient care, what evidence should we require before we are prepared to act on it?

Recent work from Harvard Medical School researchers got me thinking about what happens as promising AI research moves closer to patient care. I look at some of the questions around evidence, clinical use and who is responsible for making those decisions.

I don't think we need to have all of those answers while research is still developing. But as AI creates increasingly sophisticated ways of understanding disease and supporting treatment decisions, these are questions healthcare organizations will eventually have to answer.

AI governance can look very different once development begins to scale. I look at what happens when models, data and architecture continue to change while the governance process struggles to keep pace.