How AI Could Be Your New Doctor
As artificial intelligence is becoming the forefront of innovation in the scientific field, more and more traditionally human-led occupations are being challenged by the notion of being replaced by better, more efficient algorithm-led machinery. With many blue-collar jobs being overtaken by machinery over the past decade, we see the stress that technology imposes on the job security of many members of our community. However, with recent advances in military health research, along with an ambitious Defense Joint Artificial Intelligence Chief, even the most prestigious positions in society are being considered for replacement, in this case, doctors.
Dr.Tetteh, the Chief of the JIAC who is leading this charge, has extensive experience with AI in the medical field. Through nearly 30 projects using AI in detecting and diagnosing various medical conditions, such as early onset cancer and depression, Tetteh is very optimistic about the use of AI in aiding doctors’ efforts in accurate and efficient diagnosis’. Comparing the use of AI today to the use of X-rays decades ago, Tetteh hopes this newfound technology will become commonplace, and sees no reason why it should not be given the data collected already.
I found this article and related topic relevant to the probability laws we are learning in class, given that these new practices will undoubtedly need an extensive trial period in which the accuracy of the AI will be tested. Before becoming commonplace, the success rates of the AI diagnosis should certainly be recorded so that it can actually be proven to aid a doctor in their practice, and not just act as a sinkhole for medical research funding. We can apply Bayes’ Rule to this new innovation, given the importance of the accuracy of the tests being dependent on the prior event of having an AI diagnosis. In this case, testing should follow Bayes’ rule of having a probability of an event be affected by the prior knowledge of a prior related event, the accuracy of a diagnosis given that an AI did the diagnosing. As previously mentioned, I believe a major component in the success of integrating AI into the medical field depends on whether the AI is actually successful in its diagnosis and if it is actually helping doctors with their practices. If the trials result in a similar success rate for doctor-diagnosed cases compared to AI-diagnosed, then the technology has hope of becoming commonplace, however, there is more testing to be done to show that the AI is actually making a significant difference in helping doctors become more efficient.
Source: https://www.wsj.com/articles/how-antificial-intelligence-will-make-your-doctor-smarter-tetteh-medicine-military-covid-ai-11636123757?mod=Searchresults_pos5&page=1