Automation – designers’ friend or foe?
3 minutes read
In the early 19th century Britain was in the process of being transformed by the industrial revolution. Industries such as manufacturing and textiles were being transformed by innovative technology like the steam engine and the power loom. Previously home-based workers had to migrate from the countryside to newly built factories in towns and cities. Thousands found the skills they had taken years to master were no longer needed, replaced overnight by huge, hulking steam powered machines. One set of workers decided to fight back. Known as the Luddites, they took sledge hammers to the machines and destroyed them as a form of protest. The response was swift. The Luddites were brutally suppressed and their leaders either hung or sent overseas to penal colonies. The British Government introduced the death penalty for anyone found ‘machine breaking’ and with the help of the army the Luddite uprising was put down. The machines, or rather the mill owners had won.
I was reminded of the Luddites when I saw some of the responses to the recent video posted by Jordan Singer on Twitter of an AI driven Figma plugin called ‘designer. The tool uses GPT-3 (an AI driven natural language processing tool from OpenAI) to automatically create a Figma design from the following text:
An app that has a navigation bar with a camera icon, “Photos” title, and a message icon, a feed of photos with each photo having a user icon, a photo, a heart icon, and a chat bubble icon.
In the video Jordan can be seen launching the plugin from within Figma, entering the text and then hitting a rather fancifully titled ‘Design’ button. After a bit of thinking the computer comes back with a shiny new mobile app Figma design with a navigation bar, photos and the icons that Jordan asked for. It’s not exactly a design that Jony Ive (chief designer of the iPhone) would be proud of, but it’s still damn impressive stuff.
Responses to the Tweet have ranged from the tongue in cheek:
To the rather more concerned:
Clearly like the Luddite weavers, designers are not immune to the ever-marching impact of automation. Sure, it’s much easier to automate a manufacturing process than knowledge work, but if Microsoft can replace journalists with robots (AI will now be used to curate content for MSN rather than real people) then is it not only a matter of time before designers are also replaced by machines?
We are entering a new robotics and AI driven industrial revolution. Manufacturing, distribution, farming and countless other industries are increasingly becoming automated. The robots really are taking over.
Whilst it might be a while yet before a T-1000 is deployed to design the next breakthrough product, rather than hunting down and terminating John Connor, we’re certainly starting to see automation increasingly be used within the design process. A few examples include:
- Automatically creating design specifications with tools like Zeplin.
- Automatically creating working apps and websites from designs with tools like webflow and Thunkable.
- Automatically generating design alternatives (also known as, generative design) with tools like Autodesk.
- Automatically evaluating designs using multivariate testing with tools like Google Optimize.
- Automatically analysing customer feedback with sentiment analysis tools like Amazon Comprehend.
The thing is, whilst machines are great at non-creative tasks, like crunching data or turning pixels into working code, they suck at having to think for themselves. In the same way that AI written fiction isn’t going to be troubling the book charts anytime soon, we’re not going to see the creative aspects of the design process automated for a long while yet. And by creative, I don’t just mean coming up with ideas and nice visuals, but problem solving, lateral thinking and exploring potential solutions.
If you consider the now infamous British Design Council double diamond design process, the potential to automate aspects of the process tends to increase as you progress from the first to the second diamond. We’re already starting to see aspects of the develop and deliver stages being automated, such as generating design alternatives and automatically creating code from designs. Automating the discover and define stages however is a whole different ball game.
In computer science there is the well-known concept of ‘Garbage in, garbage out’. In other words, if you use garbage as your input (not literally, but in the form of flawed inputs) you’ll get garbage as your output. The concept of ‘garbage in, garbage out’ is also true of any design process. If you don’t understand the problem you’re solving and therefore introduce garbage into your design process, you’re going to get garbage out at the end. Have a read of my previous article – Why every design should start with a problem for a great example of this.
Designers first and foremost should be problem solvers and understanding hard complex problems is something that people are good at, machines are not.
Rather than the enemy, designers should see automation as a helping hand that can help them to work better, to work faster, and to work smarter. Tools like generative design, sentiment analysis and the ‘designer’ concept can allow designers to let machines do the work that they’re good at, so that the designers can get on with the work that they’re good at.
Caleb Meyer an industrial designer on the team at Autodesk who are developing the generative design technology summarised it nicely when he said:
“You’re not mastering the tool any longer, you’re mastering the problem — and letting the computer do all the work.”Caleb Meyer, Industrial Designer – Autodesk
Sure, at some point the machines will take over and we’ll be battling for our very survival (unless global warming, coronavirus or a massive asteroid has already wiped us out). Until then I’d urge you to see automation as your friend, not your foe.
- Let’s talk about that GPT-3 AI tweet that shook designers to the core (UX Collective)
- The rise of the Demigod designer (UX Collective)