The Curious Case of DeepMind, Google, and the Streams App

No, we do not have AI doctors yet. What we do have, however, is a group of companies that are hard at work getting us to that point and beyond.

Simply due to Google’s 2018 takeover of the DeepMind company, it would appear that DeepMind’s healthcare and medicine wing is leading the pack in this case. While this event is almost common knowledge, what most people do not know is exactly what DeepMind brings to the table related to AI in the healthcare and medicine spaces.

Because a quick Google Search on the subject yields an impressively wide range of results, focusing on one specific case of DeepMind’s foray into AI-backed medicine could be helpful.

First and most importantly, when Google decided to truly take DeepMind Health under its’ wing, both companies apparently made clear that DeepMind’s chief goal is to develop a reliable AI assistant that can augment the work of doctors and nurses. Almost right after both of these announcements were made, Google came under fire from The Guardian, which appeared to accuse it of breaching the rights of patients in its’ acquisition of DeepMind Health.

By all appearances, the primary argument here seems to have been that the acquisition proved that Google would not keep its’ promise of keeping patient data separate from Google account data. Considering this, it is reasonable to wonder how much patient data an AI assistant would need to access to efficiently assist medical professionals and how it would do so securely.

To fully understand the answer to this suggested question, it’s important to begin with what Streams, which is DeepMind Health’s AI Assistant, really is. According to The Guardian, Streams is an app that began with a simpler and much narrower focus. Reportedly, it has grown from simply being a tool to help doctors and nurses more accurately identify a certain form of kidney injury in patients to a repository for patient records and conditions related to them. If this is true, then a serious issue could arise with this sort of pivot.

As I have already suggested above, security becomes paramount when any endeavor is undertaken related to AI and healthcare, especially when this sort of effort involves locally storing patient records. Local storage on centralized servers begets easier hacks as we have seen time and time again with companies like Equifax and LinkedIn.

What makes Streams so special that it will not fall victim to the same fate?

According to DeepMind as well as Futurism, two important factors come into play here. To begin, it is backed by the National Health Service in the UK, which appears to indicate a strong level of support from the United Kingdom’s government. Only time will tell however, exactly how strong this partnership is. Secondly, on a technical level, the acquisition of DeepMind by Google means that the app will be scaled by arguably the most connected company in the world.

Who doesn’t know Google’s products?

Google’s power via name recognition alone could set Streams up to be the most successful AI app in medicine, in history. With this under consideration, the primary issue with such a partnership is not whether Streams will scale to achieving its’ ultimate goal, but if Google will make it enough a priority to truly help it get there.

A company can only spread its’ resources so thin before some of its’ investments fail. Still, working in Streams’ favor is the fact that it represents one of DeepMind’s chief achievements thus far and DeepMind essentially represents most, if not all, of Google’s AI wing.

Only time will tell how Google’s plans for Streams and DeepMind pan out, though the news cycle around the two continues to appear promising.

Furthermore, knowing these facts is half the battle. Choosing to put our trust in a system of medicine that seems bent on minimizing the involvement of the human is a different thing altogether. Moving forward, you can expect us to keep you posted on developments related to Streams and DeepMind. For now, keep in mind that the idea that AI teams are bent on completely replacing human medical professionals seems to be unfounded. Google and DeepMind merel want to ease the burden that these people face on a daily basis.

In our next piece, we’ll delve further into precisely how Streams aims to establish itself in this fashion.

Resources: 

https://www.theguardian.com/technology/2018/nov/14/google-betrays-patient-trust-deepmind-healthcare-move

https://www.theverge.com/2018/11/13/18091774/google-deepmind-health-absorbing-streams-team-ai-assistant-nurse-doctor

https://www.cnet.com/news/equifaxs-hack-one-year-later-a-look-back-at-how-it-happened-and-whats-changed/

https://deepmind.com/applied/deepmind-health/deepmind-health-faqs/

https://deepmind.com/blog/scaling-streams-google/

https://futurism.com/google-health-deepmind-streams-app

About Ian LeViness 113 Articles
Professional Writer/Teacher, dedicated to making emergent industries acceptable to the general populace