Experts predict where AI will go in 2018–and the role designers are destined to play.
Artificial intelligence was one of the hottest topics of 2017. AIs developed new languages in which to converse; the debate raged over whether AI will automate our jobs or augment our skills; and the insidious ways in which bias can infiltrate AI systems became all too real in our politics. It also become clear that designers will be at the forefront of making AI responsible and intelligible to users–whether they’re ready or not.
So what will 2018 look like, compared to the tumult of the past year? Co.Design spoke to five experts in design and artificial intelligence to find out.
New Tools Will Make AI More Accessible
Those tools wouldn’t necessarily make AI accessible to every single designer, but they’d give a more average developer who hasn’t specialized in machine learning the chance to start digging into an entirely new way of thinking about algorithms. And these tools are starting to crop up everywhere: Google’s natural language processor Parsey McParseface (yes, that’s its name), Amazon’s conversational bot builder Lex, and Microsoft’s more general AI platform Cognitive Toolkit.
For Rolston, though, the ultimate goal is for non-technical people to be able to access the technology. “Ultimately we want the Squarespace for AI, but before we get there, we need to create the Flash toolkit for AI,” Rolston says. “People who write Flash, they still call themselves programmers, but they don’t have to worry about different aspects of the programming anymore.”
Visualization Will Help Us Understand–And Trust–AI
As AI becomes more democratized, it will fall to designers to create graphics and interfaces to help people understand how the technology is working under the hood–which will play a vital role in building trust in it. In 2018, designers will experiment more and more with how to visualize a technology they’re just starting to understand themselves.
“It’s easy, when you think of AI predictive models, to create a system that’s a panopticon,” says Caroline Sinders, a machine learning designer and user researcher. “You can have the best intentions and make the worst thing. But what’s important with AI is how data visualization needs to be part of that. We’re dealing with systems where people maybe understand what the algorithm is but they don’t understand what it’s doing. Can you show or illustrate how or why the system you’re using is making the connections it’s making inside your product?”