Will artificial intelligence be, one day, powerful enough to obviate the need of randomized clinical trials?

Abstract

Artificial Intelligence (AI) currently occupies the headlines in the media as well as the medical press, with messages of its emerging unlimited potential, but also tales of doom and gloom predicting risks for the extinction of humanity! One such headline in the London TIMES of 6^th June 2023 reads: - "Two years to save the world, says AI adviser". To the uninformed, AI is nothing more than the processing of a colossal amount of data with lightning speed, now achievable with the latest spectacular developments in computing. However, could such power be harnessed to accurately predict the therapeutic potential of a new treatment emerging from Phase I, or Phase II trials, without the need to proceed to Phase III randomised, at times placebo controlled, trials? The deontological and ethical concerns about randomised trials have frequently occupied the medical literature as several clinical researchers feel uncomfortable with this kind of clinical research, especially when a placebo is allocated to patients with cancer, enlisted in a Phase III trial. Others have argued that in contrast to the belief that randomised controlled trials are more reliable estimators of the efficacy of a treatment, investigators have found that, in some settings, observational studies did not overestimate the size of the treatment effect compared with their randomised counterparts. An accurate prediction of the true potential of a novel treatment with AI, thus obviating the need of a phase III, randomised trial, could save time, effort, and finances, as well as relieve the ethical burden of allocating treatment at random to patients with cancer and limited life expectancy.