The researchers at Google Brain have recently published the results of their research on artificial intelligence in the Nature journal.
Discover our latest podcast
A system that is almost fool proof
With the help of several scientists from Stanford University, the researchers at Google Brain have been able to develop a technology that can predict the probability of a patient’s death when they are admitted into hospital. Although this medical development may seem frightening at first glance, it could actually offer additional help to medical staff.
Christened the ‘Medical Brain’, the system is capable of ‘thinking autonomously’ thanks to a network of artificial neutrons. Its mission is to analyse medical files in order to treat them as efficiently as possible. When working with 216,221 files, the system was able to predict the length of hospital stay of a patient, and even their probability of dying, according to the information it has been given. In the end, the system’s diagnoses were 95% reliable, compared to the 86% for traditional methods.
Will hospitals soon be equipped with Google devices?
For example, there was a case where a terminal cancer patient was estimated to have a 9.3% risk of dying by medical staff. On the other hand, Google’s Artificial Intelligence estimated a 19.9% risk. A few days after arriving at the hospital, the patient unfortunately passed away, as predicated by the calculations of the Medical Brain.
This should be enough to encourage the Mountain View company to continue developing these tools, to be able to offer them to hospitals around the world very soon.
