A Future of Design Without Designers?

Back in 2014 a start-up called The Grid launched a crowdfunding campaign for the world's first AI driven site builder. The vision was simple. The tool would analyse your business to create a unique design built around your exact needs. It was bold. It was future thinking. It was riding the crest of the Machine Learning and automation wave. And it was ultimately destined to fail, leaving lots of disgruntled backers out of pocket. 

While the idea was exciting, it didn't surprise many designers. They’d spent their whole careers mastering design and knew it was a lot more complicated than coming up with some randomly generated interfaces. The UI was just one part. There was a whole lot more to design hidden below the waves. Something that machines would probably never be able to do.

This is of course a common refrain from professionals so close to the complexities and benefits of their discipline that they can’t imagine it ever being disrupted. Everybody from manufacturing to retail thought their field was impregnable to technological shifts, until it wasn’t. This is especially ironic considering how many designers work in the tech sector, yet struggle to imagine it disrupting their daily profession. As a result, The Grid has become the design equivalent of Boo.com. Proof, if anybody needed it, that the industry is impervious to change.

And yet Boo.com turned out not to be the proof every bricks-and-mortar retailer hoped it would be, that people would never buy clothes online. Instead it was just a few years too early, so I wonder if the story of The Grid may be offering designers a similar sense of false hope.

If you look around the tech industry at the moment, you see a couple of competing trends. On one side of the equation you’ll see the rise of component libraries like UIKit and Tailwind UI, which allow developers to throw together relatively convincing looking interfaces with little or no design capability. On the other side of the equation you have increasingly sophisticated no-code tools like Webflow, which purport to do away with the need for a developer. While both these approaches will only get you so far at the moment, the quality and customizability threshold for all these tools continues to grow.

Somewhere in the middle we have a bunch of new workflow tools. Tools which allow you to update your React components in a graphical UI and have them automatically deploy to your live site with no need for developers, or for developers to update components and have these changes reflected across all your Figma files. We then have cool little proofs of concept which show people assembling interfaces from pre-set components using voice or text interfaces. 

At this stage, most of these experiments are a bit of fun to drive social media buzz. However it’s not going to be too long before designers have clippy like agents helping them along the way. “It looks like you’re designing a check-out process. Would you like me to add the credit card transaction pattern?”

While the area of “Generative Design” may be new to digital designers, architects, aerospace engineers and physical product designers have been exploring this space for a while. In the physical product world, designs can now be automatically generated by a series of algorithms to provide an optimal level of performance. The resulting designs often have weird, almost alien shapes. Designs that almost certainly couldn't have been imagined by traditional designers, but were instead the result of generative algorithms. 

A few years back I heard about a designer at Nike who had written a set of generative algorithms to create custom performance shoes for the USA olympics team. I was so enamoured by the story that I asked him to speak at our UX London event. What Lysandre outlined was a future where fashion designers essentially became curators. Looking through hundreds and hundreds of auto-generated designs in order to pick the ones they felt would capture shoppers' imaginations. The idea of the designer as curator rather than maker has stuck with me ever since. 

In a digital realm, the performance elements that could feed into algorithmic designs wouldn’t be things like surface friction, humidity and the gait of particular athletes, but rather click-through rates, basket size and lifetime value. Companies like Booking.com and Amazon lead the way in multivariate testing, so it doesn’t feel like a huge leap to imagine a time when the test candidates were created algorithmically rather than by human hands. 

I imagine this would start with micro copy and simple interface elements. Let’s cycle through hundreds of variations of this button style with 50 shades of blue, in order to see which one performs best. However once you start chaining these tests together, it doesn’t take too much imagination to picture whole user flows constantly redesigning themselves for optimal gain. The UX equivalent of algorithmic trading. A whole digital design team working 24 hours a day, constantly wringing marginal value out of an already heavily optimised system. Hillman Curtis outlined as much in his seminal book about Digital Design. Of course we’ll still have the local maxima problem. But that’s a good, and potentially cost effective, problem to have. 

This brings up an interesting question around the future of design. Now I’m not saying what I’m about to outline here is a likely or desirable scenario, but it is within the realms of possibility. A future where engineers, executives and PMs make the bulk of the conceptual decisions because the barriers to entry when it comes to making design decisions—in the form of craft skills—have all but been removed. A lot of designers like to talk about a world where “everybody is a designer”. There’s a chance that agentive technology may make this happen sooner than we all expected, and in ways that weren’t necessarily inline with our personal beliefs and values. 

The natural end point of this line of thinking will be designers taking a much smaller role in decision making. Essentially being consigned to creating and maintaining design systems. Some designers already feel this is happening, with PMs making the bulk of the conceptual decisions, and I know I’m seeing more and more start-ups launch their first product without a designer in sight. 

I think we’re a long way off from fully automated luxury design, so I don’t think we’ll see the end of Product Designers anytime soon. However it’s an interesting potential direction of travel, and one I’ll be watching closely.