YT randomly recommended me this video, probably because I am developing an AI platform myself, but it is very interesting we can already use an AI in clinical trials. I think the problems you have mentioned are mostly details that will be fixed with a small improvement, like biases, hallucinations, etc.. but what we really need is for AIs to have actual reasoning skills, planning, executing, long-term thinking, etc.. those are some of the last missing pieces. After that we can do something much closer to real digital-twins. And with a perfect digital-twins you could probably skip the whole clinical trials as such and release the drug, quite literally to the public without waiting for years. But that is still a sci-fi for now.. and still might take about 2-5 years until we get to that point unfortunately. And then probably another 10 years until it is gonna be approved by the slow regulatory system lol.
Thanks for such insightful webinar.
YT randomly recommended me this video, probably because I am developing an AI platform myself, but it is very interesting we can already use an AI in clinical trials. I think the problems you have mentioned are mostly details that will be fixed with a small improvement, like biases, hallucinations, etc.. but what we really need is for AIs to have actual reasoning skills, planning, executing, long-term thinking, etc.. those are some of the last missing pieces. After that we can do something much closer to real digital-twins. And with a perfect digital-twins you could probably skip the whole clinical trials as such and release the drug, quite literally to the public without waiting for years. But that is still a sci-fi for now.. and still might take about 2-5 years until we get to that point unfortunately. And then probably another 10 years until it is gonna be approved by the slow regulatory system lol.