Google is continuing to penetrate the healthcare space with its improving artificial intelligence (AI) designs. The latest in the company’s series of projects is a more advanced version of its AI that can predict the outcome of patients’ hospital visits.

The algorithms used by the new AI are based on neural networks, which are built on anonymised medical information taken from over 200,000 patients over a period of 11 years. These algorithms can predict whether the patient is most likely to be discharged, readmitted, or die.

In the US, there are about 100,000 deaths and expenses of almost $20bn per year as a result of unplanned hospitals visits. If Google’s algorithm can warn doctors to act sooner, this could significantly reduce overall healthcare costs and increase patient survival rates.

Google’s new AI is another step in the new era of cognitive computing and digital disruption in healthcare. As seen with many of Google’s other ventures into the healthcare space, this newest idea sounds very promising; however, it is possible that this new technology will face many challenges as this is still uncharted territory.

AI challenges

Biology and human health are not always as predictable as we would like them to be simply because of the unimaginable number of factors that play a role within a living creature. The sheer amount of data that is necessary to understand all the biological processes in a human body, let alone predict them, is still far beyond the current capabilities of this newest AI.

Intuition and human creativity will still play a big role in the projects where AI is applied to the human health, as no AI platform is anywhere close to replacing humans in these cases.

Another challenge with the advancing AI presence in the healthcare space is security. As with any type of big data storage, there are always concerns about attacks by hackers, as has happened before.

However, with the accurate, standardised and continuously updated input of big data into a secure database, as well as constant improvements in the AI algorithms combined with a touch of human creativity and intuition, GlobalData expects that physicians and patients will highly benefit from these developments. Eventually, the survival rates, particularly in chronic disease, should increase, and the overall healthcare cost should be lower. However, it must also be remembered that the cost of safely storing this copious and ever increasing amount of data is yet to be seen, and might negatively affect the growth of AI use in the healthcare space.