5 things to avoid when designing for AI

Tove Blomgren, Doberman’s Creative Director, was recently at Nordic.design talking about what to avoid when designing for AI.

Nordic.design is an international conference all about design. Doberman’s Tove Blomgren was there talking about AI, and more specifically what to avoid when designing for it.

The 5 things to avoid when designing for AI

Don’t try to be human

When designing for AI, don’t set too high expectations on the behaviour of the end result. Today, with the current technology being so far from human intelligence, all we’re going to achieve AI-wise is a really dumb person. Instead of trying to create a smart human, instead try to create a smart machine.

Does your interface really need a sassy personality?

Think of what personality you want your machine to have, if any. As humans, we already project personalities on things, we see stuff in things that are not there. Instead of adding personality to an item, add more value, and the personality will come with it.

Don’t confuse CUI with AI chatbots

A conversational UI and AI is not necessarily the same thing. CUI’s have been around for a long time, Microsofts Clippy is a good example. He was around before there where even talk of AI. It’s not like chatbots aren’t going to be good for anything, if it’s the right kind of interface, then it’s the right kind of interface. Examples of good applications is customer support and lawyer chatbots.

Dont’t hide the seams, hand over the power to the user and be transparent

The seamlessness between a chatbot and a person is something we’ve been trained to do as designers, but that’s something we have to reevaluate with AI. Customers might feel uneasy when they don’t know if they’re chatting to a human or a machine. Seamlessness is not always going to be good when you have machine learning in the background. The biggest reason behind that is the difference between AI and normal interfaces, it’s unpredictable. We don’t know what’s going to happen. The computer is learning and seeing patterns that we can’t. We have to assume what the experience is going to be. One example is Microsofts Tay.AI, which went from a cheery millenial to passive aggressive and straight out racist in less than 24 hours. Microsoft did not see that coming.

Don’t use it just because you can

It’s going to be really tempting to use all the new AI-fused tech currently on the market and on the horizon. Just because you could, doesn’t mean you should. Will the AI functionality distract people’s attention, focus and make their work even harder? Will it just lead to more noise?

Wherever you want to take the future or however you think it’s going to play out, it’s going to come at us really fast. The development of new tech is exponential, which might seem scary, but the cool part is that it’s up to us. We can help shape what happens.

So think about what you want to do. If we have intent and purpose then we can help shape the future together.

Check out the full presentation here.

Read more