Google Medical Brain AI Can Predict When Patients Will Die With Alarming Accuracy

Lest there be any doubt to Google's mindset as a company, Google CEO Sundar Pichai recently said that artificial intelligence is more important than electricity or fire.Google also recently rebranded its Google Research division 'Google AI' to better reflect its commitment to one of the biggest fields in technology at the moment. AI development has seemingly endless implications, and one that Google is particularly interested in applies to healthcare. Specifically, the company's Medical Brain team developed an algorithm that can assess health risks, including how likely a patient is to die.

This is one of the biggest challenges in healthcare at the moment—traversing through mountains of data, analyzing it, and making it all presentable. It's a costly thing to go through dozens, hundreds, or even thousands of patient records, and also incredibly time consuming. Google's system is able to do it, however, and can even look at notes in PDFs or things that have been written down on old charts, then analyze the data and make fast and more accurate predictions.

In one example of how Google's algorithm proved more accurate, a woman with breast cancer and fluid in her lungs met with two doctors and underwent a radiology scan. Computers at the hospital assessed her vital signs and records, and determined this had a 9.3 percent of dying. Google's algorithm also assessed the woman, but looked at more than 175,000 data points and came and increase the chance of death to 19.9 percent. A few days later, she died.

That's a frightening thing to know, that a piece of technology says you have a nearly 20 percent change of dying, and that the computer model is pretty accurate. Beyond the morbid nature of it, however, is something that could potentially save lives. It's not about predicting death in and of itself—it's about painting a better picture of a person's health, based on their data records, vitals, and other information that can be plugged into the system.

On the business side, this is the thing that could finally get Google a major foothold in the healthcare industry, while simultaneously boosting the street cred of its AI technology (not that it needs more street cred).

"They’ve finally found a new application for AI that has commercial promise," a Googler told Bloomberg. It falls right in line with Google considering itself an AI-first company, and gets Google to a place that co-founders Larry Page and Sergey Brin have repeatedly tried to grow the company's presence.

Of course, there are privacy and ethics concerns with a company like Google being able to sift through medical records and make predictions about a person's health. Navigating that could prove more tricky than the technology itself.