The system was highly accurate and one day may assist doctors in diagnosing complex or rare conditions, a Nature Medicine study found

There have been a plethora of movies and television shows illustrating the dangers of artificial intelligence (AI). (My personal favorites are the series, Battlestar Gallactica, and the movie Ex Machina). Physicist Stephen Hawking also warned about the dangers of AI. “The development of full artificial intelligence could spell the end of the human race,” the late scientist famously told the BBC in a 2014 interview. 

Yet there are plenty of reasons to embrace AI and the tool appears on its way to becoming almost conventional in research. Robots are helping streamline the process of drug development. Three facilities in Charles River — Reno, Nevada and Den Bosch, The Netherlands–have successfully integrated automation into multiple bioassay workflows, freeing researchers from considerable ‘grunt’ work. Eureka also recently looked at how deep learning can help with drug discovery: Virtual screening can computationally sift through vast collections of molecules in search of those that might exhibit a particular biological activity and quantitative structure-activity relationship (QSAR) modelling attempts to establish a mathematical relationship between chemical structures and biological activity data in order to predict the activities of molecules that have not yet been tested experimentally.

The clinic is also a ripe area for AI research. A study reported on Tuesday in the New York Times, noted that doctors were able to accurately diagnose common childhood illnesses, from the flu to meningitis, after processing the patient’s symptoms, history, lab results and other clinical data. The findings, reported this week in Nature Medicine, relied on the records of nearly 600,000 Chinese patients who had visited a pediatric hospital over an 18-month period, according to a Feb. 13 article in the New York Times. The system was highly accurate, the researchers said, and one day may assist doctors in diagnosing complex or rare conditions, according to The Times article.

In the study, trained physicians annotated the hospital records, adding labels that identified information related to certain medical conditions. The system then analyzed the labeled data and a neural network was given new information, including a patient’s symptoms as determined during a physical examination. Before long it was able to make connections on its own between written records and observed symptoms. In fact, the software was more than 90 percent accurate at diagnosing asthma. The accuracy of physicians in the study ranged from 80 to 94 percent, according to The Times.

Still, whether such a tool would ever be practical in the US is a matter of debate. The centralization of medical records in China means that it is able to amass the right kinds of clinical data for a machine to make a correct diagnosis. In the US, despite inroads in to electronic record-keeping, data is still scattered at multiple sites and systems. and privacy laws make it difficult to access data.