BEACON TRANSCRIPT – Google’s efforts to use machine learning and artificial intelligence in health care appear to have paid off. In a recent report, Google declared that its latest AI algorithm could make accurate medical predictions such as time spent in the hospital or the likelihood of death.
AI Algorithm Capable of Culling Through Patient’s Records in the Blink of an Eye
Using artificial intelligence to assist doctors in planning a patient’s treatment or to analyze disease patterns is not something new under the sun. Quite recently, one of Google AI’s peers was more successful in diagnosing skin cancer (melanoma) compared to human doctors. What AI lacks in experience makes up in processing speed.
Recently, Google’s AI and machine learning division has successfully tested a neural network-powered artificial intelligence capable of making health prediction based on the patient’s medical records.
The team said that the AI algorithm could cull through hundreds, if not thousands, of pages to make its predictions. More than that it can read scribbles at the foot of the page.
To see just how advanced Google’s AI algorithm is, here’s what happened in a hospital that accepted to test the algorithm. According to the case report, a woman with terminal lung cancer was admitted to the hospital last week.
Moreover, doctors said her lungs were filled with liquid. Based on her vitals and bloodwork, the human doctors surmised that her likelihood of death was 9.3 percent.
After feeding the same data to Google’s AI algorithm, the likelihood increased from 9.3 percent to a staggering 19.9 percent. Regrettably, the patient passed away a few days later.
The test performed by Google’s AI division is the first of many. So far, Google’s machine learning researchers agree that some sectors, such as healthcare, could benefit from having a powerful data processing unit.
Image source: MaxPixel