Can machines design?

We inhabit a world where robots have became everyday companions. Most of them are invisible entities operating hidden infrastructures and decision making, some others are on the surface and express themselves through visual interfaces, few of them, maybe the sexiest, are embodied artifacts, androids and mechanical oddities.

We want robots to work for us, carrying out the most boring and repetitive assignments as well as the dangerous and risky jobs, but also performing beyond human means; faster, smarter, stronger.

For longtime, machines have been relegated for tasks which can be split into single and repetitive acts. Design, as well as other humanistic professions, is much more complex than that. As said by Skaggs in Meta Design with Machine Learning How can we — and should we — program the deep semiotic net we take for granted into a machine, along with the symbols that are packed into every artifact in a given culture, that fluid network of signs that shifts with fashion and generations, through space and time?.

However, the set of developments known as Artificial Intelligence are allowing machines to perform activities we thought too human, — like composing music, writing poetry or even dreaming — it seems is time for robots to become creative.

So what about design?

When we wonder if a robot will design we face two kinds of questionings.

The first one is on the nature of design; can we describe design as a linear, discretional, rationalized process, more like a line, ready to be assembled by a robot?.

The second is about the capability of certain mechanisms to perform complex and contextual behaviors; could a machine be sensitive to blurry requirements that implies multiple, social biased stakeholders?.

For long time, thinking about design as a creative profession implied that we wouldn’t be able to understand the process underneath the outcome. That was the genius artist myth. Nowadays, approaches like design thinking, data driven design and others, are moving the practice into a new arena, one in which we get closer to a methodology. Design is the outcome of a process.

In this tone we find cases like The grid, a service introduced as an artificial intelligence websites builder. The algorithm behind the operator applies color palettes, adjust typography to maximize legibility and crops images detecting faces and framing the best shot. It automatizes the craft of the design process.

Other case is the Adobe’s machine learning program that takes a large body of design output from professional designers for conducting machine learning. Maybe in the near future Adobe not only licenses the software but the designer too.

The aftermath of designing machines

So, it looks not so far the big hype of robot designers. And yes, we will embrace them, but also we should take them with caution and prudence.

When we trust robots the decisions of appearance, content and even functionality of our designs, most of the time we will be sentencing our outcomes to a lukewarm homogeneity. By the end, machines will focus on the statistical median where the average user inhabits, and that’s not so bad, but we shouldn’t forget that disruption and innovation relies on the fringe and as long as machines keep iterating all designs will look and work alike.

On the other hand, the arrival of robot designers pushes our reflection of what is design and what’s our value as human designers. In the best of scenarios that kind of conversation should lead us to better practices.

Here I leave these concerns on how different stages of the design process are affected by the interference of artificial agents.

Execution and craft

The designer was born from the artisan tradition. With the strengthen of industrial tools and specialized software, the focus of design displaced from the craft to the space of concepts and ideas.

Expression

What font shall we use? Which color palette is the most accurate? What kind of feelings does it stimulates on the user? What will he interprets? How will he react?

In design schools we learnt semiotics as the instrument to reflect on the cultural interpretation of the message. But now, metrics and real time analytics are the way we validate and inform those kind of decisions. Some years ago designer Douglas Bowman wrote about that in his farewell letter to Google, but every digital designer knows about the pressure for validating each minuscule design decision with testing and evidence.

Ideas

The idea is the core of the design solution. It’s related to the problem we need to solve, the situation we want to improve, as well as the value proposition; the greater why. It’s hard to believe that a robot could take our place in that delicate and confusing stage of the design process. Nevertheless, we can find tools and aids, like the Design Thinking Canvas Autonomous. And who knows what’s coming next…

Conclusion

In recent decades we’ve seen several professions and skills disappear from the market. Maybe in the near future the role of designer as we know it will be threatened. However, it’s not about thinking how will we be replaced, but how far could we get when we got robots tackling the most boring tasks of the process. This is not to fear but to push the boundaries, to reinvent ourselves, as designers and as humans.