Artificial intelligence solutions can only become stronger and better if they get access to new data. Unfortunately, it appears some AI companies are inadvertently given access to information they should never access in the first place. Google’s DeepMind has accessed UK medical data by accident, even though it has a partnership with the UK’s National Health Service.
DeepMind May Know Too Much
A leaked letter from the UK government’s data advisor claims DeepMind has gained access to NHS patient data. It appears about 1.6 million people have been affected as a result. This information was shared in order to train a new medical app. It now appears the NHS – and Google’s DeepMind- are facing additional scrutiny due to this information changing hands on an “inappropriate legal basis.”
As is to be expected from such a government communication, there will be an official investigation regarding this matter. This deal between the NHS and DeepMind has been causing a fair bit of controversy for quite some time now. It is expected a conclusion of the investigation is imminent, although no one ventures to guess how things will play out in the end.
To be more specific, there is a lingering question of whether or not DeepMind has the legal authority to handle NHS patient records. Every single record contains a lot of sensitive information regarding the individual patient. Moreover, the AI company accessed this information without asking patient permission, which could prove to be a major problem. That all depends on how this investigation turns out, though.
The fact this information is used to train a new medical app based on artificial intelligence will not help matters either. According to DeepMind, the company accessed this information to “provide direct care,” yet that will be a topic of substantial debate for months to come. Training a new app or providing direct care are two very different things, even though the app is used by doctors and nurses across the UK already.
More specifically, Streams – as this app is called – is used to identify patients who may suffer from acute kidney injury. It is one of the main causes of death in the UK as far as medical conditions are concerned. In fact, it appears about 25% of these deaths can be prevented, assuming medical personnel can treat patients at an early stage. To do so, they need new tools and equipment to help determine if the patient suffers from acute kidney injury or not.
It will be interesting to see if there will be some sort of repercussion for either NHS or DeepMind regarding this “data leak.” Exposing confidential patient information without consent is not a smart decision by any means. Although Google’s AI subsidiary will claim the information has not been sued for commercial purposes, it remains to be seen if that is indeed the case.
If you liked this article, follow us on Twitter @themerklenews and make sure to subscribe to our newsletter to receive the latest bitcoin, cryptocurrency, and technology news.