Bob Newhart in the World of Blade Runner
And other adventures in AI “art.”
Like everyone else working in content and media, I’ve been messing around with AI art and text. There are some thorny questions around automating human effort the ethics around how the machine was trained. Others have weighed in so I won’t here.
I was mostly staying away, and then a VXF supervisor friend convinced me to dig in a little more. He uses Midjourney to make quick look books and rough in tone/vibe. This makes a lot of sense to me. With the carefully considered prompts, you can get to fairly specific looks more quickly than you could searching for swipe and/or sketching (especially you’re me and you haven’t really been cultivating your drafting skills lately).
For pre-concept art tasks, I’m sold. Putting together quick comps to demonstrate how an idea might work visually or to get cross functional teams aligned? Yes. For finished work? Not so much.
This is partly due to the aforementioned thorny issues, but mostly because I’m noticing AI generated art (at least, Midjourney) seems to all have a look. I can’t quite pin it down, because stylistically it can be all over the map from photorealistic to illustrated to aping specific artists, but every time AI art shows up in my various social feeds, I instantly “make” it as AI and it bugs me.
It’s not just the freaky fingers that AI hasn’t worked out how to deal with. Maybe it’s an artifact of the algorithm. Or maybe it’s been trained with too small of a data set. (I have no idea how this actually works.) But there’s a subtle patina that seems to imbue everything.
Now, I’m too old to get away with being a young Turk blowing up the status quo, instead risking relegation to “old man shouts at cloud” status, so I’m conscious of needing to be open minded with emergent technologies, and I’m happy to have my mind changed. And in the meantime, I will be using it in supporting conceptual tasks instead of, or in additional to, image searching (and my aforementioned feeble sketching skills).