AI Needs UX Designers, Too

“Agentive technology” is a quietly humming buzz-phrase, appearing in my Twitter feed from time to time (and now my physical library at home), and it always piques my interest. The idea of a very focused artificial intelligence that works in the background, trying to make my life easier without making a spectacle is, frankly, the coolest. But it’s still got a long way to go.

I have friends who have a Nest thermostat in their apartment. A few weeks back—while in the middle of reading Chris Noessel’s book about agentive technology—I was over at their place and noticed the little smart thermostat on their wall. It’s round and slightly bulbous and, dare I say, a little charming. Its purpose is to learn your schedule and temperature preferences, and then create the perfect environment for you without your intervention. Having never really lived with one, I just assumed it worked well. So I asked my friend about it.

“Oh, we just use it as a normal thermostat,” he explained with a wave of his hand, delineating all the ways their Nest had failed them, even after customization and “training” the device. It just never worked as well as they’d hoped. My “smart tech” infatuation immediately triggered the “oh, you guys probably did something wrong” response in my head, which is, of course, wrong: as soon as you start blaming the user, you shut your eyes to the problem. It’s not the user. It’s the tech.

The Nest thermostat is an example of what Noessel refers to as “agentive technology.” Alongside the Nest are a host of other smart devices and software that aim to learn user preferences and environmental stimuli and adjust accordingly. It’s part of the idea behind the Internet of Things (does anyone still use that term?), a bunch of smart devices all around you that talk to each other and do things on your behalf so you don’t have to. We’ve been promised the Internet of Things since as early as 1982, this future where technology figures out on its own how to make our lives simpler.

But it’s not quite here yet. Because in order to work, this technology has to be seamless. It has to understand what you want it to do, what you tell it to, and how to improve to make your life even easier. And even more importantly, in order to promote adoption, it’s critical to ensure we’re using this technology to solve the right problems, at the right time, in the right way. People will often try new technologies; if they’re not impressed (or worse, put off), the technology can die.

Before I ramble on more about all this, quick note: If you’re at all interested in designing for these types of applications, I strongly recommend reading that book I’ve now mentioned twice: Designing Agentive Technology: AI that Works for People by Chris Noessel. You can’t Google “agentive tech” without seeing his name. So if you’re interested, read it.

Anyway. My point: this technology has long been driven by engineers focused on the actual AI side of the equation. When the technology is at the center of the design process, the user is not.

Agentive technology isn’t going anywhere. Our technology has been increasingly “smarter” as the years go by, and it will continue to be smarter whether UX designers get more involved or not. So like last week, where I rambled on and on about how the burgeoning field of consumer-friendly virtual reality needs more user experience designers, I’ll urge UX designers here, in general, to learn the technologies that are on the horizon. At the very least, you’ll have cool topics to talk about with your friends. At the most, you’re making the world a more user-friendly place. This technology has a lot of potential to free up cognitive space for our users, allowing them to focus on the things that matter most to them. That’s no small thing.

And even more importantly, write about your experiences and lessons learned with designing and developing for these technologies. We all benefit from learning from the community. So please share your experiences! I’d be remiss as a UX Booth editor if I didn’t ask you to share them with us, of course.

If you’re looking to learn more about all this, Noessel’s book is definitely the place to start. Here are a few other articles to check out too, most of which cover AI a bit more broadly, but include the more focused stuff referred to as agentive technology:

• The Adobe blog, on how AI will affect UX
• Great UX for AI
• AI Won’t Change Companies Without Great UX
• AI First—With UX

The folks at Rosenfeld Media were awesome enough to send us a coupon code for those of you who want to check out Designing Agentive Technology: AI that Works for People! Just head over to their site and enter the code uxbooth17 at checkout for 20% off.