It is no secret that healthcare sector holds a humongous amount of data. What if this data is fed to an artificial intelligence (AI) to foresee the judgment day of a patient? Recently, researchers published a paper in Nature, describing how AI can accurately predict the last day of patients by analyzing the electronic health records (EHR). During trials, researchers used EHR from two U.S. hospitals. Researchers fed the data to an algorithm to predict the duration of a patient in hospitals, and even their time of death. The data included patient’s medical history, vitals, and even the handwritten notes of doctors. The algorithm creates patient’s timeline and predicts the length of patient’s stay in a hospital and takes one step further to predict the death time. However, what is the use of this algorithm beside scaring patients by telling their doomsday?
The technology’s primary aim is to improve patient’s care by adjusting the treatment plans and foresee any medical emergency to avoid health hazards. Moreover, the technology can reduce the workload of healthcare workers, who spend hours in converting data into a standard format. However, researchers believe that the biggest challenge is the collection of such massive data of each patient when most of it lacks the standard and semantic interoperability. The assistance of AI can reduce all the mundane tasks of collection and interpretation of reports and charts. For instance, a study showed that if large data regarding eye diseases and retinal images fed to an AI, it can predict the odds of a patient developing an eye disease. Moreover, a recent AI-based technology can detect lung cancer and heart diseases better and more accurate than doctors.
Even though such innovations dazzle people with their abilities and advantages, there is a hidden factor that they forget to mention. To make such predictions, AI will need a massive amount of centralized data. Google already knows too much about a person’s lifestyle, habits, and choices. Giving access of healthcare data to Google can lead to chaos if someone tries to misuse the information. What if Google shared this data with third parties just like Facebook, what will happen then? Recently, the U.K government raise eyebrows by accusing Google that its technology DeepMind Health collected users’ information without their consent.
There is one more factor that needs to be addressed. AI gives outcomes are based on the data fed to it and can be manipulated by feeding biased data. Recently, MIT researchers developed an AI that thinks only of murders and death, just because it was fed with only Reddit contents. What if Google’s DeepMind Health AI also gets manipulated, who will control its repercussions? Therefore, to prevent the misuse of healthcare AI, we need a regulatory framework, which can ensure transparency between people and AI-based technology. It will be hard to enforce regulations on private companies such as Google, but if want AI to be a boon to mankind, we need to create proper protocols to control them.