Trite phrase, accurate this time.
I've played with several of the varied AI art tools. Me? Well, I get bored easily, and, since I have SOME talent, I'm not overly interested in playing more with the tools.
Every time new tech has appeared certain segments of the population have gone doom and gloom - in the 19th century photography was supposed to be the death of art - what's the point of painting and drawing id anyone with a box can just capture the scene? Of course the photographer had to know how to frame an image, work his camera and deal with the chemistry of development. Still work of a human artist.
Hell, the CG revolution of the 1990's was supposed to be the "end of art," since the computer just "did everything."
Of course the worries then were hyperbole. In 3D computer graphics a human artist still had to build the models, paint the textures, animate the motions, light the scene, set the camera angles, etc. Another human artist still had to composite the passes, add in other elements and manually twiddle controls to achieve a final result. Even generative filters still require the human artist to adjust settings.
The new slate of AI generative tools is fundamentally different. In most, you type in a text prompt, maybe, "a company of medieval underground delvers unloads their caravan in preparation before entering the entrance of a cave to the underworld..." And the computer will just scan the stored database of billions of images and start generating images for you. No muss, no fuss. Don't like what the computer spits out? Hit reprocess and it'll generate more. Want to change the image? Well, you can change the prompt. You have to tell the computer the cave entrance is a 200 foot high giant skull. But, other than telling the computer what to generate from a verbal prompt, you have no control over the output. (OK, some tools have a "style" setting which forces it to look like whatever images the database has tagged as that style.) It really IS the computer generating everything.
It won't kill art. You need a human to reproduce specific designs (if you tell the computer to create an Imperial Stormtrooper, you won't get Star Wars armor - but you'll get something close, and the AI tools can't do Era Ten EE suits - yes, I have tried that as a prompt - because the AI hasn't ever seen any of the EE Suits Red designed, Todd refined, I altered, and the artist who did the Ground Zero cover re-interpreted. Professional, storytelling art - comics, movies/TV, etc - will still need human artists for now to all take direction, stay on the same page, and be consistent. You still need a human artist to interpret and implement specific guidance (since the AI is highly random).
But the AI is potentially devastating to the starting artists, concept artists and artists who generate work for micro/small budget work. Above my example prompt was, more or less, a brief for the cover of Better Games "Designing Underworlds." Now, I know the game, the genre and the author, so Red mostly left me alone to create and play and gave me some feedback notes as I iterated my design, but, for the most part it was my call. I chose the 3D models (created by other talented artists) and photo references (one from Pexels, one was my photo). I had to pose then place the models, etc... It took me a few days and, for a "working artist," the amazingly small pay made it "not worth the time." When I DO charge for a project, well, Red paid me for 30 minutes of work... This cover was a fun project for a friend and I'd have done it for free.
Now Red doesn't have to scour websites looking for clip art. Red can blip over to Midjourney or Dall-E or Wombo, type in his prompt, let the computer do its thing and have several dozen copyright-free images in minutes.
Since we all know small game publishers have almost no budget, its in the interests of those small publishers to use AI tools. You'll get more polished looking material in less time than your buddy who can draw a bit can produce for $20-$50 bucks an image.
There's AI Rotoscoping now. An entire specialty in the VFX world is obsolete, and the roto specialists are all frantically trying to gain other skills because their jobs are going away. This is not hyperbole. I have friends who are active at multiple studios and each one of them just canned most of their roto staff. Hundreds of artists lost their jobs to an AI.
Of course technology has replaced people in jobs for centuries. This isn't entirely new, but it's a huge shake up for the art world.
Below is a totally safe link. It's my created file in my Dropbox account. It's over 40 images I generated on my phone while typing this (and after typing this), using the Wombo Dream app on my phone. Now, Wombo tends to spit out more abstract images than something like Midjourney or Dall-E, yet can absolutely see these being used in game books as interior art, or even a cover. I spent more time waiting for ads to play (what? Like I paid for Wombo Premium?) than I did for the AI to create anything. This doesn't include all the ones I threw away instead of saving. These 55 (saved) images were generated (along with probably another 50 I threw away) in just over an hour, including the ZIP and upload to Dropbox time, mostly while sitting on the crapper. Obviously, with a few more Prompts and another hour or two I could have the AI generate enough art to fill out multiple game rulebooks. I could even cough up the $4.99/month to not have to sit through ads, access other processing styles, download higher resolutions and get faster image generation through the bot. So, yeah, for $5, and with an hour or two of time I can generate hundreds of usable images, many of which are better than any single image that would take me days to create, and way faster than it would take to browse stock sites for free images. And, again, Wambo Dream isn't even considered one of the good tools.
So, yeah, working artists are scared shitless right now. They should be. They're about to get a lot less work.
And Red - these are AI generated images, so I have so sense of pride, ownership or copyright. May as well add these to your stock image library, cuz I can absolutely see many of these appearing in future games. Ok, USMC at Waterloo has sailed.