Well, AI is here, so what do we think? So much has been said already by so many others, and to be honest, I’m still wrapping my head around the implications already. To be specific, when I say AI, I’m referring to the broad concepts of text-to-image Generative AI stuff like Midjourney and Dall-E and Stable Diffusion. Pick your flavor, they all kinda do the same stuff right now, and riff on the overall concept of being trained on previously existing artwork being scraped en masse from the internet, then that data is more or less digested, and from inferences made from the data, they can accomplish a sort of effective “autocomplete” for images, and come up with some crazy stuff, some of which is laughable, but some actually compelling as well, at least at first blush. Is it plagiarization? Is it “just a tool”? Is Art DOOMED???
Basically, it’s too soon to tell. Hype cycles aside, I’ll admit this stuff is advancing in output quality very quickly and there’s legit cause for alarm, especially for character illustrators, background painters, and concept artists, and broadly for designers in general. There’s a good case to be made that these programs will put artists in these fields under great stress- in fact I’m certain it’s already happening, from discussions that I have with colleagues, and the new reality of watching clients opt for trying AI solutions where previously there was no option but to hire a human. But, it might just be temporary- there’s this nasty little thing called Copyright Law, and currently (its June 2024 at the moment) all the legal rulings coming from the Copyright Office are pretty clear, generative AI outputs are not copyrightable, and they really aren’t wavering from that position much that I can tell. Ultimately, I predict, the Copyright office will outright deny the prospect of any Generative AI outputs as copyrightable materials, and I think this might help curb the appeal of those looking to cash in on creating intellectual properties with these shortcuts. I’m old enough to remember to rise- and then fall- of Napster. The laws will catch up with this stuff and my bet is that it won’t go great for the Tech bros.
Additionally, and this is also based on stuff I’m seeing out there in the news and in real jobs- there’s just a huge public hazard involved with letting this tech run off unimpeded- and politicians will act (slowly, but they’ll get there) to make sure that generative AI outputs are forcibly labeled as such- if for no other reason than to prohibit disinformation sabotaging the political processes. Political deepfakes are already a thing and there’s gonna be a lot of personal incentive to protect themselves from those obvious threats; and by extension, recognize the greater damage unimpeded AI generation can wreak on the news, education, and criminality at large. So eventually, media created via GenAI will require a little visible “AI” stamp, akin to ®, ™, and © symbols we all recognize.
Enforcement of that won’t be easy, but eventually I predict GenAI will split into two distinct camps- those in compliance (opting to label their outputs as AI and obey related legislation and terms of service), and opposite that, those driven underground, who will basically trade in disinformation, disruption, and scams. At that point, for the greater public, we’ll probably have as much appreciation for what AI can do for us as we have for spam emails and robocalls and junk mail. Consumers, who, over time, will each have their own “I fell for an AI scam and got burned”, will then utterly reject most AI output as scammy, cheap, and basically tainted. Today, AI is a hot buzzword, but like NFTs and Blockchain and Metaverse before it, it will, I predict, inevitably become cringe.
It’s definitely accelerating the “enshittification” of the internet, basically rendering previously reliable image searches questionable and sowing confusion on social media platforms as to what is or isn’t accurately depicted- and coinciding now that the public is pulling back from social media in general. As if the Bots and Ads weren’t bad enough, adding GenAI to it is basically gonna be the final nail in the coffin- nobody wants this.
Even further, Adobe (makers of photoshop) and social platforms from Meta, X, reddit and even FUCKING WORDPRESS have updated their terms of service to include us basically opting in by default to having our content scanned and trained on for purposes of feeding their AI machines. Uproar is rampant in the artistic communities regarding these developments and everywhere I look my colleagues are sharing info on how to mitigate this exposure. A new artist social network called Cara has gained momentum seemingly overnight, promising digital autonomy and privacy from these scraping tools. While I’m a bit skeptical, I signed up an account for myself just to park my handle and if nothing else, help send a message. But it’s really hard to pick a path to commit to when so much is in upheaval. Where is our new artist sanctuary? And will our audiences follow us if we move?
Privacy should important to corporations too- I’ll concede, it’s a hot thing now to see how AI solutions are able to save time and money, but is that worth opening up corporate secrets to a black box entity? How can you be sure the GenAI is maintaining security? There’s no way to hold a program liable to an NDA. I think eventually, at least as an official stance, corporate lawyers will squash the temptation to save a buck by skirting security protocols. I could be wrong there though, if “secure AI” becomes a thing.
And while some have embraced the potential, it seems the vast majority of creative workers are fighting back against the tide. Artists have testified before congress and worked with developers to create anti-scraping image protections like Glaze and Nightshade, and while new, they are potentially able to draw a line in the sand and deter further general media scraping. I was also heartened to see the WGA and SAG have held their ground enough to stave off the AI takeover of their professions, at least for a while. I hope this holds true for the IATSE during this year’s negotiations, and especially my union, the ADG, where AI is definitely an existential threat for many members.
Overall, I can foresee an eventually swing away from the current fascination with AI- but how long might that take, and what damage happens in the interim, is very hard to say. Despite AI being a robust image generator, and GenAI outputs finding their way into my client pitches, I’m happy to say I’m still getting hired for storyboarding projects- even for commercials that hype the latest upcoming AI solutions.
I have, I can admit, experimented with AI a tiny bit. I wanted to assess the threat for myself, after all. I asked ChatGPT whether AI could replace storyboard artists, and it replied that, basically, it probably wouldn’t. I’ve also tried asking it to generate storyboards of simple scenes (like a father teaching his son to shave, as evidenced above), and it’s not that great at it at all- at least for now. Granted, if you ask for a photoreal image, it does an amazing job. So I dunno, I can’t quite explain why asking it for a black and white sketch is harder to achieve than a photoreal image, but it just is. Good for me, I guess.
At one point, I put a board sample into one GenAI app and had it attempt to do a paintover- the results were pretty impressive- but i was horrified to see that the AI put another artist’s signature (or facsimile thereof) in the corner of the output. So much for that use case.
While I definitely predict more pain in store for the creative class in the short term, I think ultimately there are enough reasons to be optimistic in the long term. I had paused for a long while with updating my website, fearing what might happen if my works get scraped and if I get “trained,” and yes, that could happen, but hiding isn’t going to help either. So I’m going to continue putting out my best work and make it clear that there’s a distinct advantage to hiring the best human artists, and hope that clients will appreciate the difference.
I’m also trying out something new- it’s a bit of a new venture for me, and I’ll probably suck at it for a bit, but I’m giving a go at doing Art Streams on Twitch. The reasoning is partly to give me something to occupy my downtime, and make friends, broaden my networking, plant seeds for a new kind of audience I hope to cultivate, and perhaps most importantly, shed some light on the HUMAN creative process, and set a good example for young artists who might otherwise be tempted by the Dark Side of AI shortcuts.
Regardless of whether AI stays or goes, I’m going to focus on consistency, fundamentals, and customer service- especially my discretion- and always on bringing my perspective on mastering traditional and new drawing methods to quickly achieve my client’s needs, and making sure value and experience is always evident in my drawings.
Additionally, from here on out, I’m Glazing/Nightshading every damn thing I post, anywhere. Here’s hoping it makes a difference.
Thanks for reading!