Disclaimer: Opinions expressed below belong solely to the author
Before you ask, yes, the characters you see in the featured image above were not created by humans.
As a graphic designer myself, I’ve always been rather skeptical of artificial intelligence (AI) taking design jobs from people. The obvious argument has always been that creative work is — or should be, at least — original, while AI can only learn from what has already been done.
For instance, you can feed the machine learning algorithms millions of images of past work and they can procedurally learn how to mimic them, but shouldn’t really be able to create something new.
However, I was just reminded that isn’t necessarily true. And I should have known better, since it applies not only to machines, but humans as well.
You see, the thing is that most creative work is also derivative. We learn from others and then incorporate their styles and techniques into our work. Historically speaking, there have been very few revolutionary breakthroughs that changed everything — be it in design, art, music, writing or any other creative pursuit.
Quentin Tarantino once asked about his success and he replied, “I didn’t go to film school, I went to films”.
This is the same principle on the basis of which new AI solutions are sprouting. Some of them are so advanced that they appear to be laying ground for entire new industries.
Design in seconds
What if, instead of hiring a designer, briefing him on the project and waiting days or weeks for drafts before moving to subsequent stages, you could just type in your request into a website and receive your designs within seconds?
Logos? Done. Photos? Right here. A fancy painting? Wait just a second. Icons for your app? No problem, here they are. A fancy cartoon character? Check this one out. Everything within reach without Photoshop or Illustrator, without even hiring a human being.
Well, it’s already happening.
These instructions — that you use in recent viral AI engines like DALL-E (developed by OpenAI, an independent research lab co-founded i.a. by Elon Musk) or the more dreamy, artsy Midjourney — are called ‘prompts’ and there are already people specialising themselves in writing them, before selling them on recently-launched marketplaces like PromptBase (launched in June), where anybody can buy them for US$2 to US$5.
All you do then is go to the AI solution they were written for, paste it, and wait for the algorithm to produce the image. Voilà!
While it’s a nascent platform, some people are already cashing in on it, though it will take some time before new fortunes are made there. Nevertheless, a designer interviewed by The Verge revealed he is making between three and five sales per day out of a pool of 50 prompts that he keeps adding to.
That’s already at least a few hundred dollars in passive income that is bound to grow with time as more people become aware of the service (as ever — the early bird gets the worm).
But why do we even need prompts?
Right now, finding what can work is a bit of a trial and error and both DALL-E and Midjourney have limitations. The former is still in closed Beta, while the latter requires a paid subscription after the first 25 creations.
Most of the instructions you enter may not produce the desired effect yet, and the pretty, professional images are typically a result of a fairly lengthy process that requires tweaking. It takes repeated attempts, which cost money — starting from US$10 for 200 “jobs” on Midjourney, and US$15 for 460 images on DALL-E, to get what you want.
Which is why it makes sense to pay a few bucks for a prompt that will produce a quality result.
Everything you see here was designed entirely by algorithms, following the specific prompt and you can buy each of them for your own use.
Pretty much any creative task can be accomplished with a simple instruction instead of weeks or months of planning, and without human involvement.
Just a few days ago, the news broke that an AI-generated piece won an actual art award in August, without the jury being even aware of its origins.
Mind you, we are only at the very beginning of this revolution. As time goes by, machine learning algorithms will be able of even more complex tasks and even more precise outcomes.
Are creatives going to lose their jobs?
Whenever someone is making money using a novel technology, it typically comes at the expense of a profession that is being displaced by the innovation. So, yes, many designers, musicians, artists or even writers, may soon be forced to find a new job.
But, as it has happens with every such revolution in history, the best will stay in business, and perhaps even earn a premium.
Mass production of furniture a’la IKEA did not make all carpenters go extinct. An electronic watch did not diminish the value of a Patek-Philippe. Cheap, mass-produced fast food did not make restaurateurs bankrupt. And while most people may depend on the productivity of mechanised farming, the market for small, organic agriculture is booming.
Hence, as ever, many people are going to be displaced by the innovation that answers the vast majority of the market’s needs.
But as it does that, it’s going to a) create plenty of new jobs for people specialising in handling AI, and b) the market is bound to reward the best human creators more handsomely than before.
(Oh, and let’s not forget about c) it may actually help existing designers in prototyping work, early design stages, providing far more inspiration than otherwise could be generated, accelerating projects which typically took months.)
Interestingly, it may soon impact hiring as well. Unlike soft skills, teamwork and communication with other humans — which have been emphasised by companies looking for new hires in recent years — the most rewarding and sought after competences may soon be in knowing how to speak to… machines.
Featured Image Credit: Midjourney, Community Showcase