The Low Expectation of AI in Design
Artificial intelligence has received a lot of attention recently with the public availability of ChatGPT text demonstrations. I’ve also seen a visual version able to create simulated images based on existing photos. Some have told me that they are scared that design might disappear because of this technology. I don’t agree.
I think this enthusiasm for AI in design, writing and photography comes from two false premises: ideas are rare and the idea is the hard part.
First, AI is good at generating a lot of ideas. Some of them, I’m sure, are useful and innovative. AI probably is faster than a human at this aspect of the design process. From what I’ve seen, it would take me a week to create as many ideas (at the same resolution of detail) as photo creation AI seems to do in minutes. I can see a use for this in design: input the project constraints (ie the design guide / aesthetic concept) and then review the results and pick the most appropriate designs for developing more fully.
However, there are already a lot of designs and ideas available. Most of them are free and instantly available on image search and apps like Instagram. There are hundreds of designers and illustrators whose entire careers are built on promoting their abilities by posting brilliant ideas on the Internet. Almost none are patented or could be lightly modified to avoid infringement.
Another way of imagining the glut of ideas we have is to consider creative writing. London in the late 1500s had Shakespeare, Marlowe, Ben Jonson and many other of English’s finest writers. The population of London at that time was about 200,000 people, or the size of Topeka, Kansas. If a city of 200,000 created a dozen brilliant writers, imagine how many must exist in a country like the US with 330 million!
Also, any store is filled with ideas that are available for anyone to build on. For example, I recently got a Dyson vacuum. Much has been written about Dyson’s patent assets, but there are also hundreds of little mechanisms and details on the Dyson that could be used on other products or even on competitors vacuums. I really love the hose clips for example.
So, if we already have a large quantity of ideas, why aren’t we using them? Because development and distribution is the hard part.
In books, this means having editors who can make sure a text is ready for print, publicists who can generate interest in reading them, graphic designers who can lay them out, publishers who can print them, a supply chain to put them in stores and cashiers to put them on shelves and check out customers. Writing is a huge personal achievement, but in the grand story of a book, It’s a small part of the process.
In design, this gets even smaller. GM has 157,000 employees, but only 1,900 designers. When one takes into account all of the work that is outsourced (from manufacturing to dealer sales) the ratio of designers to employees is probably far lower. I think the ratio is largely similar at places that I’ve worked.
Getting in even closer, I would say that less than 10% of my time is generating initial ideas. 90% of what I do is research, CAD, reviewing with engineers, sending drawings out for quote and assembling prototypes.
Lastly, is the human element. We desire not just objects and text, but human stories. Doubt this? Look at all the fake people we already have invented to sell product: Uncle Ben, The Quaker Oats guy, Tony the Tiger, hell even Thomas Edison and Elon Musk are as much fiction as fact. A product connects with us in a deeper way when we feel the emotion of the creators. AI can fake that, but there will always be a difference.
