biloxi: Thanks for posting this FDA PR regarding its intent to "integrate" AI into its drug approval process in an effort "to radically increase efficiency in bringing new medicines to market." As someone who has an intense interest in a particular drug being approved as soon as possible,I welcome any reasonable effort to shorten the mountain of man hours that has historically driven the regulatory review process. And if, as stated in the PR, AI will also be utilized to "Predict Drug Efficacy and Safety," a concept similar to the use of surrogate endpoints to predict efficacy in the accelerated approval process, I'm all in on that as well. In fact, even though I am a novice in AI, I wouldn't be shocked if a properly programmed AI determined that the 5 Leronlimab Ladies are not an accident of nature and that women dying of TNBC should have the opportunity use it.
Yes, I fully understand that the utilization of AI for this purpose, even with honest intentions to fulfill the FDA's regulatory mission, will inevitably produce mistakes of omission and commission that will need to be rectified. But the current system of evaluating new drugs is such a mess, on so many levels, that attempting to harness AI to create efficiencies in that process strikes me as a very worthwhile endeavor.