There is an over-promise of AI in all fields of medicine, to be honest. I think this is driven by major geopolitical changes and huge hype of the technological companies and the rapid increase in size of the data centers available for foundational models to be trained, etc., so I see a lot of hype everywhere, to be honest. I think the areas that are most likely to have an impact on the treatment of, or diagnosis of, our patients are probably the applications of foundational language models, like ChatGPT-style tools for patients...
There is an over-promise of AI in all fields of medicine, to be honest. I think this is driven by major geopolitical changes and huge hype of the technological companies and the rapid increase in size of the data centers available for foundational models to be trained, etc., so I see a lot of hype everywhere, to be honest. I think the areas that are most likely to have an impact on the treatment of, or diagnosis of, our patients are probably the applications of foundational language models, like ChatGPT-style tools for patients. And I also envisage a lot of applications in imaging, for example, for bone marrow imaging, et cetera. To be honest, most of them are being developed, but not yet approved for patient treatment. So that’s why I said there is a lot of hype, because many of these technologies are being developed – they have very promising data, but are still immature to be implemented in real daily care. So I think globally that could be the area that’s probably going to have more impact in the short term. Then you have a lot of very interesting fields of work, and I think it’s going to probably deliver a lot of promise in several areas, like, for example, clinical trials. It’s an ongoing discussion about whether we adopt or not synthetic branches in trials, and this could speed up a lot of research, etc., but still, it’s not a reality in this sense, and it’s not AI – but there has been, strictly not AI, the FDA has recently approved kind of recommendations for the acceptance of Bayesian statistics based on Bayes’ theorem, which will probably change the way we do a lot of trials – but strictly, still not AI, although probably advanced statistics will probably make a fast change in how we do trials in the forthcoming years.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
