Google has an AI algorithm that uses your eyes to predict heart disease

AI is the technological acronym of the moment, appearing tied to everything from smart speakers to smartphones. Scientists at Google’s health tech arm, Verily, however, have started using artificial intelligence to assess a person’s risk of developing heart disease.

Verily has developed an AI algorithm which is able to analyze scans of the back of a person’s eye and determine their age, blood pressure and whether or not they’re a smoker. Basically, all of the potential cardiovascular risk areas. Using this information, it’s then possible to assess someone’s risk of suffering a heart attack or developing heart disease. 

The researchers trained the algorithm using machine learning by having it analyze the eye scans and general medical data of around 300,000 patients. The AI was then able to gather patterns from this data and understand which signs in the eyes scans pointed towards cardiovascular risk.


In a paper published in the Nature journal Biomedical Engineering, the project researchers said that the method would have to be subject to further testing before it can be used in clinical settings. 

However, it has the potential to be just as accurate as current methods of assessment while being faster, easier and far less invasive. In the testing phases, the algorithm was able to identify heart conditions 70% of the time, which is a slightly lower rate of success than the longer SCORE process which is correct around 72% of the time.

While it might seem strange to look into a person’s eyes to determine their heart health, it’s actually not unusual and the eyes are actually a good place to check for the first signs of many health conditions. When it comes to the heart, the back inner eyeball is covered in blood vessels and their appearance can give all manner of insights into heart health. 

Essentially Google has taken a diagnosis method with an established history, found new ways to analyze the data and sped it up significantly. It’s clear from its Project Baseline that Verily sees great potential in continuing to gather medical data for analysis and it’s likely we could see a great deal more innovation in the coming years. 

Emma Boyle

Emma Boyle is TechRadar’s ex-Gaming Editor, and is now a content developer and freelance journalist. She has written for magazines and websites including T3, Stuff and The Independent. Emma currently works as a Content Developer in Edinburgh.