How can Microsoft balance the benefits of AI and pervasive computing with user privacy? Do they even need to?

3

I’ve been thinking about a couple of interesting articles over the past few days, namely Paul’s ‘The Failure of Windows Phone and the Next Big Thing’ (https://www.thurrott.com/mobile/windows-phone/82610/failure-windows-phone-next-big-thing) and Brad’s ‘Microsoft’s Home Hub Looks To Bring Cortana To The Kitchen’ (https://www.thurrott.com/hardware/83654/microsofts-home-hub-looks-bring-cortana-kitchen). 

I currently have no interest in talking to my computer. I made a concerted effort when Windows 10 came out bundled with Cortana to use it as much as I could and make it a part of my day to day workflows, but ultimately it’s still just easier for me to use my trusty mouse and keyboard. 

That being said, I think AI and relatedly pervasive computing are going to be big. While hovercars and holodecks are still a distant dream, the sci-fi future where I can walk up to my house, unlock the door using Windows Hello, have the lights come on and my favourite music automatically start playing when sensors detect it’s me who has walked into a room, ask Cortana what food is in the fridge and have one of the Echo like AI enabled devices located in every room respond to me, tell it to play a movie on one of the various connected screens located throughout the house while I do some cooking, ask it to order me a self driving taxi to arrive so I can make a date in 30 minutes depending on the traffic conditions, leave the house and have my movie playing where it left off in the car is just around the corner, if not here already. Assuming this is an area MS does decide to get inolved in, of course, although this is appearing increasinly likely. 

While I think this sounds amazing, it would also mean that Microsoft, a for-profit company, would know almost everything about me, and this makes me somewhat uncomfortable. Knowledge of movements, tastes, habits, communications, far greater than they have today. And of course if Microsoft have it, what guarantees are there that it wont be accessed by a government agency? Or released in a data breach? 

But does this even matter? Having a camera on a laptop pointing at our face all day and carrying around a GPS tracker in our phones everywhere we go are things that 30 years ago would sound uncomfortably 1984ish, but we now demand that our devices have these things, and base our buying habits around them. Conversely, there were real concerns – even outcries – about the privacy implications of Google Glass. Windows 10 perhaps fell somewhere in the middle, with complaints about data tracking and the like being common in the early days but quickly dying off. 

Where will future forays into these fields fall? Will privacy concerns prevent AI like Cortana from becoming truly mainstream, and keep it relegated to an interesting novelty? What about devices like Home Hub? Will we feel uncomfortable allowing them more and more access into our homes and lives, controlling appliances, lighting, heating and communications, instead keeping it as a glorified voice enabled MP3 player? Or will these concerns be swept aside as we embrace the utility of the devices. 

Talking specifically about Microsoft, if they were to get involved, what guarantees could they make, or what systems could them implement to reassure people that only necessary data is being gathered, and that it is being kept safe? Cortana’s Notebook feature seems to be a good step towards this. And could Microsoft leverage it’s image as a security conscious safe pair of hands (Brad has an interesting article about this here: https://www.thurrott.com/windows/windows-10/83177/microsoft-security-company-premium) to get an advantage over the likes of Google who are traditionally seen as only wanting users so they can commercialize their data, or Amazon who presumably just want to sell us as much as possible?

Would love to hear your thoughts. 

Post Reply

Join the discussion!

BECOME A THURROTT MEMBER:

Don't have a login but want to join the conversation? Become a Thurrott Premium or Basic User to participate

Register
3 Comments
Sort by Votes | Date
  1. 0 | Reply
    TenFour - 1 month ago

    Good comments. It is not a matter of "if" all this data being gathered on our personal lives will be released in data breaches, but "when." The only hope for most of us is to take reasonable precautions and then trust that we will be the needles in the haystack that don't get found. Also, I have no interest in having AI do all these things for me. If everyone goes to the Italian restaurant rated 4-5 stars the others will be gone. If everyone takes the cheapest flight, it will quickly become the most expensive. If everyone travels to the same "great" destinations they will be ruined. Is there really any benefit to having your door unlocked by your phone instead of using a key? Or being able to turn on my lights with an app instead of using the switch? I don't want my fridge ordering more food when I run low, because I like to shop for different things and change things up in our meals. I don't really want to talk to my computer at work, which is already too noisy, and I have no interest in talking to my living room.

  2. 0 | Reply
    jimchamplin - 1 month ago

    It's a tough question honestly. The truth is that if we want the "Star Trek" future, then our systems will need access to massive amounts of data not only about us, but about other people, and detailed facts at that. Until real-time access to information across multiple systems worldwide is available, it's going to remain "less than perfect."

    While it may sound unfair to say that it has to be "perfect," the truth is that if it isn't perfect, you'll grow quickly frustrated with it, since it'll be 90% capable, but then fail you that one time that matters. Maybe your accent isn't easy for it to understand, or it doesn't have access to something and that causes it to fail in a strange way. Get into a series of requests where all you get is "Unable to comply," and you'll turn it off and not use it again.

    1. 0 | Reply
      TenFour - 1 month ago
      In reply to jimchamplin:

      I keep reading that despite the current failings of AI "it will get better," which I haven't seen much evidence of. You still get bad directions, dangerous in some cases, from your driving app. I was recently directed down a one-way street the wrong way. You still fail to get information when you need it--trying to catch a flight and instead the phone is insisting you shop at some nearby store. This highlights one of the failings of the current implementation of AI--it is all about getting you to spend money on something. If instead you want to just take a nice walk it is hopeless.