This may change at some point of course, but here in 2017 I reckon the table below is a good reminder of where to spend our time as designers.
|Things Humans are a best at||Things Computers are best at|
|Unstructured problem solving||Calculations|
|Defining what counts as relevant||Pattern matching|
|Remembering qualities||Remembering linked values|
|Dealing with incommensurables||Predictable procedures|
|Identifying fair value-exchanges||Following simple rulesets|
|Listening / empathy -> to tailor approach||Learning from big data -> to tailor approach|
Until we’ve exhausted the design challenges in that right hand column, let’s stop thinking we can muscle in on what humans do best.
The most important work is to spot those moments when users are being forced to do things that computers are good at – like remembering to check particular things or doing little calculations in their head. If we can do that for them, they’ll have more time and energy to do the bit that they’re good at (the stuff in the left hand column).
Algorithmic AIs are assistants. We need them to handle the boring stuff. The filing, the reminders, the pattern matching, the calculations, the transactional stuff… To surface up the data we are likely to need – but with enough humility to remember that humans will always need a role in the ongoing process of determining our needs.
Ultimately this is down to the fact that we’re uniquely wired to read our context. The human soul is not a database of fielded values. We are spiritual, emotional beings with the ability to creatively understand the dynamics of our situation. Our model-making / creative interpretation has a dominant role in determining the experiences we have.
As designers of computer – human interactions, we have to remember when it’s good to get out of the way of human – human interaction. Let’s book the appointment, help surface the right product / service info to prepare for it – but then allow most of the difficult stuff to happen with an expert human.
Financial advice is a good example. We don’t need a questionnaire to understand our appetite to risk as much as we need a skilled human being to talk to. Crisis handling is another good one. We’re worried, distressed, bothered. The last thing we want is to be forced to speak to a computer.
In these cases, we’re better off looking at the systems and screens that staff use. By providing mediated access to data and functions we can better ensure that the nuances that humans can detect aren’t missed in the service interaction. Back office systems typically provide a mine-field of design challenges that fall in that right hand column.
Let’s commission those projects. Less sexy. Less innovative. Less futuristic maybe. But still a great deal more value to be gained.