Creativity on Autopilot? The AI Era in Design
AI, AI, AI — that’s all anyone can seem to talk about these days. You’d think suddenly there was a groundswell of support for the underappreciated Spielberg/Haley Joel Osment film, “A.I. Artificial Intelligence” (2001). And although I am not holding my breath that Rotten Tomatoes will re-evaluate its erroneous 76% rating for the film anytime soon, the unending buzz around AI begs that seemingly undying question: is AI taking over our lives?
Perhaps the topic is inescapable right now but thankfully, we’re not quite at world domination.
Enter, Photoshop’s Generative Fill
Last month Adobe Photoshop joined the artificial intelligence race with the introduction of its AI Generative Fill in the beta version of the software. Some might say, “Don’t call it a revolution,” but I’m a big fan of hyperbole, so I’m going to call it a revolution.
With a simple prompt, or no prompt at all (depending on your goal), you can make that landscape photo you love — the one that, while beautiful, doesn’t fit quite right on Instagram — and make it portrait orientation within less than 15 seconds. Generative Fill ingeniously fills gaps and eradicates unwanted elements from images. Its intelligent algorithms work swiftly, saving a lot of time, and crucially, saving its users a great deal of painstaking effort.
Photoshop isn’t doing anything it couldn’t do before. In fact, it’s doing exactly what everyone who’s never used it thought it was capable of for years — “Can you replace my cousin with my brother in this picture? And put him in a suit? And change the background from day to night? I need it by this afternoon. Thanks!” — but now it is doing so in seconds. What used to take even a skilled user hours to do, can now be accomplished with the click of a button.
Now, can it be done by a layperson? Sure.
Can it be done well? It depends.
Is Generative Fill the future of photography or web design (or any other arm of the creative industry that uses Photoshop heavily)? I’m not even going to pretend it could be — it’s simply not.
The Bad News and The Good News, Too
Despite its efficiency, Generative Fill occasionally produces artifacts or distortions in complex scenes with intricate details, which then necessitates manual adjustments. While anecdotal, I’ve used the tool and some results have been, in a word, unsettling.
I recently input a real picture of a person into Photoshop, intending to remove an unwanted item next to them via Generative Fill. Instead, I was subjected to the visage of what was allegedly a human, seemingly created by compiling a collage of various body parts found on the web, a result no experienced Photoshop user would actually execute.
Although Generative Fill is still in beta, Midjourney, the “Coolest Kid in the Artificial Intelligence Art World” TM (trademark by me), is already running a completely different game.
An innovative amalgamation of traditional editing methods and AI-powered adjustments, Midjourney has unlocked new dimensions of creative possibilities. By offering real-time previews of various editing options, it empowers users to make nuanced and controlled adjustments.
Yet, even Midjourney is not without issues. Mastering Midjourney’s features might prove challenging for those accustomed to conventional editing techniques. The risk of over-editing and resulting in unnatural-looking images, similar to Generative Fill, is still a concern, and a frequent one, in my experience. However, the potential for simplifying complex processes and democratizing advanced editing for novices is truly transformative.
Transformation Remains Human-Dependent
I truly believe these AI accomplishments are transformative. Yet what they ultimately highlight is the opposite of what you might expect: they show how absolutely necessary people are to the creative field.
One could enter any prompt they want and get something good enough — but for good work, great work, you need a person to determine and shape that result and bring the intended product to fruition. You need people to make the source material that AI pulls from. You need people to push the boundaries of “what is” into “what could be.”
That’s how we got here.
A host of creatives got together and created tools like Photoshop’s Generative Fill, Midjourney, and other tools that now resource the wealth of human knowledge available on the internet.
The realm of image editing has been irrevocably altered by the introduction of advanced technology and software enhancements. These features have introduced unprecedented efficiency, automation and control for artists and designers. But while reveling in the advantages, it is vital to recognize the potential drawbacks and navigate them accordingly.
And to not forget who got us here: We did.
And there’s no replacing us.
Albert Burks IV is a Senior Visual Designer at Mediacurrent. He is a wealth of useless, yet informative, knowledge about all of your favorite movies and TV shows (that he’s never watched personally).