I thought AI would replace Photoshop. Here’s why I still do it myself

Where generative AI pictures shine

As I’m not a Photoshop professional, my abilities are positively restricted. There are issues I simply can’t do as a result of I don’t know tips on how to do them, and AI is sweet for these bits. AI can also be good for fast little duties the place I don’t look after excellent outcomes, like prototypes and ideations.

Generative AI is improbable for shortly whipping up idea artwork and creating fun digital props for a roleplaying game. I’ve used it to create phony sci-fi pill overlays for sending faux messages to my tabletop RPG gamers, and Photoshop’s personal Generative Fill characteristic is a quick and dirty healing brush/clone tool replacement that makes it method simpler to fill in any gaps in a picture or simply make it a bit of greater.

I’ve used generative AI to create prototype card layouts for a sport I’m designing, for fast private memes between buddies, and for portraits representing the characters I wish to roleplay as.

But for me? That’s the place generative AI’s usefulness ends. I don’t use it to create sprawling vistas or gigantic artworks. Why would I? Sure, it may be spectacular from a technical standpoint that AI instruments can create these issues out of skinny air. But I don’t actually have any use for that.

At their core, giant language mannequin AIs simply aren’t able to understanding something meaningfully. Even once I do want generative AI, the dearth of accuracy, precision, verisimilitude, and skill to observe particular directions kills its usefulness. It nonetheless principally appears like a tech demo, and that makes it largely ineffective for something past novelty.

The evident weaknesses of generative AI

One time, I used to be making a personality portraint for one in every of my gamers in an upcoming Alien RPG tabletop sport. I wished a sci-fi man in a jumpsuit with company vibes and to have his fingers in “W” and “Y” shapes to symbolize his loyalty to Weyland-Yutani Corporation (his employer).

I struggled rather a lot with that one. I imply, I simply wished the character to have his fingers splayed in the best method. But might AI do it? Oh boy, might it not. No matter how laborious I attempted to finagle it, the outcomes sucked.

I attempted upwards of 10 totally different prompts to get it to grasp that I wished three fingers up on one hand and two on the opposite, splayed to create the impressions of “W” and “Y” letters. Sometimes it made the arms face the improper method. It by no means bought the variety of fingers proper, and it by no means splayed them in the best method. It completely failed.

After making an attempt—and failing—to get the AI producing what I wished for near 20 minutes, I gave up and simply made it myself utilizing the primary generated picture as a reference. I cloned one of many fingers, moved into the best spot, adjusted the lighting, blended the layers, and it got here out nice. All of that took 5 minutes.

Okay, high quality. You would possibly say that I constructed upon the unique creation put forth by the AI. Yes, I’m glad it gave me that preliminary design to work with, and it was simpler to edit that than create the identical factor from scratch. But as a closing product? It did not do what I wanted. It was no extra usable than a random picture I might’ve grabbed from someplace on-line.

More usually than not, it’s simply faster for me to make edits manually. With AI, I’ve to consider tips on how to immediate the AI accurately and ensuring the AI will interpret my directions the precise method I intend to get what I want from it—not one thing else, not one thing that will get it proper in a single space however randomly adjustments one thing in one other space, not the best picture however with an overhauled artwork type. With the newest crop of LLMs, doing this proves frustratingly and opaquely tough.

And that’s my major downside with at present’s generative AI: it takes a lot time to repair what it creates that I would as effectively have accomplished it myself within the first place. Just have a look at the infamous Coca-Cola Christmas ad that got here out atrocious but price way more to create than in the event that they’d simply paid some animators and artists—they usually ended up doing that anyway as a result of the AI outcomes sucked and wanted to be punched up.

Now, when you’re a person and also you don’t know the very first thing about Photoshop, then AI can be a tool to get you halfway there. But we’re not but on the level the place handbook changes, punch-ups, and handcrafted artwork—even by amateurs like myself—are out of date.

Faster, cheaper, extra sustainable

Indeed, there are one million fast little issues I nonetheless use Photoshop for the place it solely takes me a couple of seconds to do. It doesn’t make sense to make use of AI for these items, even when AI is able to them: resizing pictures, adjusting lighting, tweaking distinction, reframing a picture, changing to a distinct file kind, altering facet ratios, and many others.

Jon Martindale / Foundry

These are all necessary duties I do each single day, and there’s no method I’m going to 1) belief an unreliable AI to do what I want accomplished accurately, or 2) waste all that GPU time, electrical energy, and water on one thing I can do quicker and extra successfully with current software program. (Yes, the environmental costs of generative AI are frighteningly costly.)

Look, I’m not anti-AI. It’s not like I hope the know-how dies, and it’s not like I can’t see the way it may be helpful. But it’s necessary to be aware of what we’re utilizing and the way we’re utilizing it, and I feel it’s pointless to make use of generative AI for something I can do myself, particularly if I can do it higher, quicker, and with much less frustration.

The better of each worlds?

There’s in all probability an finish sport someday within the subsequent decade or two the place generative AI will be capable to do what I do effectively sufficient… and I would find yourself dropping my area of interest altogether. Some even assume that’ll occur to everybody and we’ll need to deal with a post-work world. But I don’t assume LLMs are what’s going to make that occur.

For now, AI instruments can assist. They’re helpful, however they’ve limitations. I take advantage of AI for enjoyable, novelty, and unimportant stuff. I’ll spin up some pictures and movies to ship my buddies, or inspiration for our subsequent roleplaying session, or a fast idea for a inventive challenge. It’s invaluable for placeholder artwork in a work-in-progress sport design, for instance.

But in the case of something vital, something that calls for specificity, something that would land me in scorching water if it accommodates errors, something I can already do myself very quickly? I’ll simply do it myself. No method I’m going to entrust any of that to an unreliable AI that received’t take heed to directions and can as a substitute inject a bunch of its personal hallucinations. If AI goes to make my life more durable, then I’ll simply do it myself.