Discussions around artificial intelligence and machine learning have been heavily dominated by computer scientists and developers. However, if we want to deliver viable AI-powered solutions that are valued by end users, designers need to throw themselves into the game and start designing more intelligently for artificial intelligence.
Since IBM’s supercomputer Watson competed on Jeopardy and won the first place prize of $1 million in 2011, artificial intelligence has been on everyone’s lips. Stakeholders in the tech industry seem to be fighting to win the race, and almost every day new artificial intelligence-based products get shipped to market. However, when reviewing existing products powered by artificial intelligence, it becomes clear that some deliver a better and more intuitive experience than others. Along with this, recent events like the Tesla accident, where a driver was killed due to poorly designed steering wheel sensors, underline the fact that artificial intelligence and machine learning products need very careful designing.
If you tune in to Google and search for “definition of artificial intelligence” or “what is AI?”, you’ll find a great number of articles trying to define artificial intelligence. Some writers even dedicate whole articles to comparing and discussing different variations of the definition. Although a general definition doesn’t exist, I find it necessary to touch briefly upon an overall understanding of artificial intelligence to have everyone tuned in on the subject of this blog.
Artificial intelligence, also called AI, is a branch of computer science that deals with imitation of intelligent (human) behaviour. In other words, AI enables machines to perform tasks, which normally require human intelligence such as visual perception or speech recognition.
Looking a the description, it becomes clear that the arrival of AI-powered products means that we’re entering a new era where machines behave differently than we’re used to. Instead of simply performing orders, machines start to act “intelligently” and do things by themselves. Not only does this change how we interact with products, it also changes our expectations for these products. As designers, we need to understand both the technology and end users in order to deliver useful and easy-to-understand (intelligent) products.
Designers need to get involved!
For way too long (if you ask me), the field of artificial intelligence and machine learning has been dominated by computer scientists and developers. But now it’s time for change! We need designers to throw themselves onto the tech-scene and take on an active role. And not only do we need designers to participate in the professional discussion, they also have to take part in the work of identifying the space of opportunity of artificial intelligence. In other words, designers ought to be involved in the decision making about where and how to apply artificial intelligence.
For this to happen, designers must get their heads around the technology and start developing a deeper understanding of artificial intelligence and how it works. In fact, this applies to all designers employed in companies dealing with new technologies — whether it’s artificial intelligence or blockchain for that matter.
Don’t get me wrong. My point isn’t that designers should turn into developers, but they need to get closer to the code. This means that designers must get involved when the programs are being written for artificial intelligence. Even in the tech industry, many designers tend to focus heavily on discovering user needs and defining concepts rather than the actual development and testing. It’s, of course, understandable since the field of artificial intelligence is complex and difficult to understand, but it cannot be an excuse to designers anymore.
Identifying the (AI) opportunity space
I mentioned earlier that existing AI-powered products leave users with very different experiences. While some products feel magical, others are unsatisfying or disappointing. So, what makes the difference?
First, simply bringing a design perspective into the process of creating artificial intelligence products will be of great value. For too long, artificial intelligence has been approached inside-out, meaning products were based on what computer scientists are (technically) capable of doing rather than what value could be delivered to end users.
When designing for artificial intelligence, designers should keep the affordances of the technology in mind and tune in their usual approach. Instead of focusing on problems or pain points, it makes sense to look for places with too much complexity for users to handle on their own. Since artificial intelligence is especially good at delegating complexity, we should let it do the heavy lifting on tasks that are either complex or repetitive and mundane. Identify the tough tasks that people can’t cope with, and start with those. An example of this is Airbnb’s smart pricing feature that combines more than 70 different factors and put forward a suggestion for your housing.
How do we humanise AI?
The important thing to consider when designing for artificial intelligence is the need for machines to understand humans. Though we still have a lot of things to learn about ourselves, we do know that a major component of what makes us humans is our ability to empathise with others. Empathy builds trust, and it makes us feel connected and shows us that someone cares. In order to make us feel connected to machines, and to feel as if machines really understand us, designers have to embed empathy in machines. But how do we make a machine care?
A way to deal with this is by exploring the underlying intentions of end users’ actions. I suggest asking the question: What would my mother do? In case you were wondering, I’m not kidding. By asking this question, you can start to explore if artificial intelligence could, in fact, deliver a similar kind of value to users. Another aspect of humanising machines is redesigning the touch point or the way we interact with (digital) products. As designers, we need to create intuitive and frictionless interactions between humans and machines. An example is products with a conversational UI, since the voice is a more natural form of human interaction. In other words, interfaces are changing from relying on technical principles like GUI to more natural human interactions.
Let’s bring AI to its full potential!
To fully exploit the potential of artificial intelligence, we need designers on board. Curious designers who want to understand the technology and how it works. Designers who are able to identify the right opportunity space for artificial intelligence, and who take an active part when the programs are being written. When this happens, we are able to humanise artificial intelligence and hence deliver products with a more natural form of interaction and machines that feels like they understand humans.
By Rie Christensen, Designer in LEO Innovation Lab