Google's New AI System Shows Promise For Diagnosing Diabetic Retinopathy

Google's New AI System Shows Promise For Diagnosing Diabetic Retinopathy

Google is stunning the world yet again. Over the past few years this tech giant has led the way in artificial intelligence (AI) research, and now a brand new AI system that Google's team has developed could prevent one of the most common diseases leading to blindness.

The team at Google has been working hard with eye doctors from around the world to understand all there is to know about diabetic retinopathy. AI specialists said they could use their AI technology, known as "deep learning," to analyze changes in a person's retinae over time.
Just in case you weren't aware, diabetic retinopathy is a complication of diabetes that often leads to blindness in people over 60. Over time, the retinae in the patient's eyes become damaged.

For all the non-techies out there, "deep learning" refers to a specialized field of AI that uses certain algorithms to help machines "think" in abstract terms. "Deep learning" technology has already proven effective at performing various highly specialized tasks. Just a few of these tasks include being able to recognize various human faces and being able to understand a person's voice.

The reason Google's battle against diabetic retinopathy is back in the news has to do with a successful study just released in the Journal of the American Medical Association. Researchers found that Google's AI was as good as well-trained ophthalmologists at recognizing the warning signs for diabetic retinopathy using patients' retinae photographs.

Google made it clear that they do not want their technology to replace doctors. In fact, Google recommends everyone get a thorough screening from an eye specialist as often as possible. Instead, Google wants this technology to help the ophthalmologists of tomorrow screen as many patients as possible.

Google is especially hoping to roll out this technology in countries where the healthcare systems are not as robust as in parts of the industrialized world. Indeed, a major impetus for this study was to try and figure out a way to help doctors in India screen cases of diabetic retinopathy earlier.

Lily Peng, who was one of the leading researchers on this project, told reporters that she hopes this technology can help doctors in her home country of India better screen patients. Peng says that Indian doctors have a difficult time taking care of all the diabetic patients in their country right now. This new AI system, Peng hopes, could really help Indian doctors diagnose diabetic retinopathy quickly and prescribe effective treatment as soon as possible.

David McColloch, who is a diabetes specialist at the University of Washington, was impressed with this study's findings. He said that if Google could devise an on-ramp for this AI device, then this technology could be used literally anywhere in the world to give quality retinal scans to patients.

While this is an amazing start, Google is on a mission to further explore the possibilities of using AI to detect other diseases. Google recently announced that their AI lab in London was working with the National Health Service in the UK to see if deep learning technology could be effectively used to diagnose other diseases.

The first warning signs of diabetic retinopathy are generally difficulty seeing colors, hazy vision, and floaters in the eyes. There are currently less than 200,000 cases of this disease in the USA.

« Back to list