The future of UI is AI

Artificial Intelligence is profoundly transforming the way we interact with digital machines.

“Her” is not a movie about a voice-controlled interface. It’s a movie about a computer being able to think and feel like a human. “Siri” isn’t cool because it talks to you, it is cool because it understands what you mean. “Hal9000” isn’t scary because it has a voice, it’s scary because it makes decisions on its own.

We are still a good bit away from having voice-command interfaces ready for actual everyday use — on the one side there are still technical glitches, on the other side there is a simple lack of social acceptance — but the layer that lays behind those interfaces is already changing the way we interact with computers on a major scale.

Artificial Intelligence sounds incredibly complicated (and real Artificial Intelligence — I’m talking neural networks, etc. — surely is), but basically it is just a bit of math. It’s the code that recognizes an appointment in the text of an e-mail and makes adding it to your calender so easy. It’s the code that looks at your Netflix history and takes a pretty good guess at what you might want to watch next. It’s the code that finds patterns in your pin boards and shows you pins you might like based on that in your feed on Pinterest. It’s the code that deciphers the gibberish on your credit card bill and sorts your payments into the right categories in your banking app. It’s the code in your thermostat that figures out when you’ll be coming home today and starts heating so you’ll arrive at a cozy home. You get the idea.

Artificial Intelligence (often referred to as “magic”) is already all around us and while we are on our way to get full-blown voice-controlled operating systems that will only become considerably more. There are many interaction designers unaware of that and this is going to become a serious problem for them: Users attention span is short. If you can’t confront them with the information they want and/or need they will be dissatisfied. You might be able to make a good guess at what that information is based on your personas, user stories and observations, but trust me: AI will be able to make a far better one. In times of increasingly tinier screens figuring out what the important part of the story is (not just some story but any story, that is) is a valuable skill. And if the result is based on your users personal preferences (that might even be unknown to themselves) that’s even better. Reading whole news articles on your Apple Watch won’t be much fun — getting the version that is boiled down to one sentence by the Summly/Yahoo algorithm will be. Getting woken up in the morning by the song you picked a year ago will no longer be an option, when you can be provided with news you are really interested in, what’s happening in your social networks right now and some new songs based on your Soundcloud likes. As you can already guess information is only one part of the story: the other one is behavior. Suggestions, notifications and choices will need to be filtered by the users previous decisions — and even less than that: by their (maybe even not interactive) actions.

No designer can consider all users previous choices and actions. Or make predictions based on their social network and previous decisions in different situations. We need to understand and be aware of the power of AI and incorporate it into our interfaces. We need to now at least a bit of how it works in order to decide what it is capable of and when to use it. We also need to know how to make it non-creepy, how to make humans either unaware of it or make them feel comfortable with it.

It feels like Artificial Intelligence has been on the verge of breakthrough for an awful lot of time. Game-AIs have been around for ages, automatic data detection is nothing new and there has been at least some kind of voice recognition years ago. But something changed in the last decade: With the (continuing) rise of ubiquitous computing suddenly there is enough data for the AIs to use to compute their predictions, there are enough users to make these predictions relevant and there is enough of a market for companies to put serious money into developing AIs and non-glitchy speech recognition.

Many jobs will be occupied by Artificial Intelligence in the near future. Let’s make sure ours won’t be part of those. If you haven’t already, you should start getting a grasp on the benefits and dangers of AI and try to think about how it can help making your users lives easier and more comfortable. Think about how to incorporate it into your interfaces today — there is no need for your users to feel like they are talking to a human being, but there definitely is need for them to feel a bit surprised and delighted by your foresight and how prevenient (in that very sense) you are — and by “you” I mean “your Artificial Intelligence”.