The Automation and Disruption of Design
Over the last five years I have been studying the development of AI/ML capabilities, and how these technologies might relate to the business of design and illustration — although I’m not an illustrator. I see a time whereby AI/ML could very possibly be disruptive to our industry; with an emphasis on the more illustrative end of the spectrum to begin with. This will not only bring into question what we do, but who gets to do what we do. This is neither a threat or a benefit, but rather what it is.
Keeping on that thought – the automation of illustration – Google just released a drawing aid they’ve creatively named “AutoDraw” (https://www.autodraw.com)— which is in practice a drawing–as–a–means–of–search tool, yet I can’t help but see AutoDraw as the seminal stages of this potential disruption.
If we extrapolate out along this trend, the idea behind LogoJoy (https://logojoy.com) is a powerful example of how AI/ML could be acutely disruptive to the vocation of designing logos and brand identities. In the last 3 months, LogoJoys revenues have jumped from $15,000 per month to $75,000 per month. LogoJoy has automated the process and delivery of logo design for thousands of customers (hundreds per week!) — this is work that might have previously gone to logo designers. I do not anticipate this trend to slow down, at all.
To support my theory further, we can look to products like Snapchat and Prisma. These products move the work away from illustrators and art workers, and instead towards complex sets of algorithms. These algorithms are available 24/7, they never get tired, they can be optimised, even hired and fired when needed, all with minimal or no involvement from conventional illustrators or art workers.
So far, I’ve focused on illustrators because it is less clear how AI/ML will affect UI/UX designers. Longterm, to remain relevant as a problem solver, one will need to position themselves further and further up the decision tree. In doing so, one will need to develop deeper and active understandings of AI/ML, psychology and philosophy. That is not to say that what we currently do becomes redundant, but rather it becomes commodified and nieche, as it meets its automation.
The good news: Someone will still need to research and design the automation goals and protocols—for a period of time, this will still require deep human involvement. In the meantime, the best thing we can do (and should be doing already) is to insist on being present at the premise of the problem, and not only at the creation of the solution. This might mean transitioning from product design into product management. Ultimately, I see this as a mandatory progression for most designers wishing to continue growing within their career over the next decade.
The automation of design will not happen overnight, but I sense, it’ll feel like it has once it does.