“Currently we pick these things up too late and harm is caused to patients, and we think there’s a real opportunity for these AI systems to be able to predict and prevent rather than just what currently happens, which is clinicians almost firefighting and running around problems that have already developed,” DeepMind clinical lead Dominic King told Wired.
The team fed health data from more than 700,000 Veterans Affairs hospital patients across the U.S. to their neural network.
Their results were promising, according to a paper about the research published Wednesday in the journal Nature: the system can even tell doctors what piece of medical data tipped it off that a kidney crisis was imminent.
But while the system is speedy, it’s way too trigger-happy: it reported two false positives for every correctly identified kidney injury.
The researchers also have yet to see how well it works with patients outside of the VA hospital system.
More on DeepMind: There’s a Power Struggle Inside Google to Control Superhuman AI