A lot of money goes into artificial intelligence research, and advocates of the technology have praised it as a way to revolutionize health care. The global market for AI in health care is expected to rise from $1.3 billion in 2019 to $10 billion by 2024, according to investment bank Morgan Stanley.
Researchers recently published new findings in the Lancet Digital Health Journal that concluded AI is on par with medical professionals in identifying diseases. However, in order to really tap into how AI can improve health care, scientists concluded more research is needed.
The research centered around something called deep learning, which uses algorithms, data and computing to emulate human intelligence. Deep learning allows computers to identify patterns of disease by examining thousands of images.
The study, led by Dr. Xiaoxuan Liu and professor Alastair Denniston, aimed to evaluate the accuracy in deep learning algorithms vs. health care professionals. The results are mixed, however, mostly because of how these studies are conducted.
Liu, Denniston and colleagues focused their research on papers published since 2012, which turned up 20,500 studies. But of those, only 14, less than 1%, provided good data.
With the data from the 14 studies, researchers found that AI and human medical professionals are identifying diseases at about the same rate. When detecting diseases, deep learning systems were accurate 87% of the time compared to 86% of medical professionals.
According to researchers there is a lot of promise in the field of AI, but the issue is there needs to be improvements in study design.
“Evidence on how AI algorithms will change patient outcomes needs to come from comparisons with alternative diagnostic tests in randomized controlled trials,” Livia Faes, from Moorfields Eye Hospital, London, said in a statement. "So far, there are hardly any such trials where diagnostic decisions made by an AI algorithm are acted upon to see what then happens to outcomes which really matter to patients, like timely treatment, time to discharge from hospital, or even survival rates."