The Dog & The Boy, and What The Advent of AI Generated Artwork Means For The Anime Industry

So I just recently got finished rewatching Watamote and I’ve been taking it easy over the last few days since writing that review—it was a bit of a difficult one to write, and barely even a day after finishing it I had an issue with my eyelid swelling which made it difficult to focus on doing much of anything. I’ve got His and Her Circumstances lined up to rewatch next (which oughta be a fucking trip) and in the meantime, I got recommended an article earlier today about a short, 3 minute long anime movie called “The Dog & The Boy”. It’s nothing particularly worthy of note to be honest; a bite-sized animation without any dialogue chronicling the story of a boy and his robot dog in some vague sci-fi setting where the two of them get separated by a sudden outbreak of war and are eventually reunited decades later. It’s not bad, but it’s hardly anything special—at worst, it’s a tiny, ultimately pointless story trying to get cheap reactions out of people through the avenue of its subject matter, the sort-of thing that makes you question “Why spend time and resources on this as opposed to, like, anything else?”

But let’s be real, we’re not here to talk about the subject matter or the narrative of The Dog & The Boy itself. We’re here to talk about something else entirely, a recent controversy reaching far beyond just the anime industry; AI generated art. I haven’t really given an opinion about the whole debacle since it really kicked off some few months ago or so, but it continues to be a hot topic and now “The Dog & The Boy” has given me a reason to discuss it in relation to the modern state of the anime industry.

I’m sure you’ve already guessed this, but The Dog & The Boy prominently features so-called “AI generated backgrounds” throughout its 3-minute long runtime. This is only partially true; in actuality, a lot of the legwork was done by real artists. A template was created by hand, then partially built upon and filled in with detail by a computer algorithm, then cleaned up and finalized by a real artist. There’s a common misconception that AI generated artwork is a complete replacement for the work of real artists, but the simple truth is that even with the technology coming as far along as it has, it’s far from being perfect, nor is it capable of replicating whatever exact vision an artist, designer or director can see in their mind’s eye. When you’re operating with a program that builds something upon the foundation of a textual prompt, you really have no way to know what it’s going to spit out; and if you have to draw a template yourself, that kind-of throws a wrench in the idea of fully autonomous artwork without a need for human input. So what exactly is wrong here? Why is using AI artwork as a supplement to or a template for real artistry a bad thing?

Firstly, in this specific case, there’s an issue with lack of credit for the artists working with the AI algorithm. The background designer for The Dog & The Boy is credited, verbatim, as “AI (+Human)”. I’m sorry, did I miss the part where you’re supposed to name the humans who actually assisted the AI in the first fucking place? You can see through the show’s own visual showcase of its AI generation process that humans were responsible for doing a significant chunk of if not the vast majority of the legwork; and even if they did only a small portion of that work, it would still be incredibly disrespectful and slimy to completely discredit and ignore the role that they played in bringing this particular show to life. In much the same way that developers are the lifeblood of the video game industry, artists and animators are the lifeblood of the anime industry. They’re the guys on the front lines doing all of the heavy lifting. In many of my reviews for shows that I dislike, I tend to make a point out of how none of what went wrong with the show was the fault of those who didn’t have any say in its creative direction and were just following marching orders from the people above them, so I find the notion that the blood, sweat and tears they shed trying to make a name for themselves in the anime industry can be so easily swept under the rug and blatantly disrespected by those trying to push the false narrative that significant chunks of their creative process can be fully automated with horribly imperfect tools downright insulting.

Speaking of those aforementioned imperfect tools and how they can supposedly hasten the process of taking something from its conceptual stages to its final ones, we need to talk about what AI art generation is actually doing. Most of the modern AI art generation algorithms out there today are “trained” through an extensive data/knowledge base of images sourced from pretty much all over the internet; just about anything publicly available for the AI to use is swiped and added to that knowledge base. And look, humans learn how to draw by studying the work of artists who came before them; that’s not the problem, because taking inspiration from something isn’t plagiarizing it. But AI is equally as stupid as it is sophisticated; it’s not capable of “taking inspiration” from something because it lacks actual creativity. What’s actually happening when an AI generates artwork—explained in the most basic and easily digestible way possible to the best of my understanding—is that images the AI recognizes as depicting specific things are plucked from its database and used as direct templates for the creation of something “new”; essentially just an elaborate fusion of already existing artwork (among other things like stock imagery and photography). Maybe some people would argue this isn’t really plagiarism because of how sophisticated the technology is and how unique the end result of that generative process can seem, but that’s missing a point. AI generation software like Midjourney is being sold to users at a premium price point; it’s technology that has been commercialized and is operating entirely upon an extensive knowledge base of artwork that has been stolen from thousands of unwilling participants across the internet.

The fact companies are actually able to get away with this without serious legal repercussions is astounding to me, but let’s set aside the legalities here and focus instead on the morality of whether or not it’s right to sell this technology in the first place. One of the most immediately obvious and common use cases of these programs is to make a piece of unique artwork imitating the style of an existing artist. As a ridiculous example, let’s say you wanted to recreate the cover artwork from Akikan’s first light novel volume in the style of legendary manga artist Kentaro Miura (who—if you aren’t immediately familiar—was responsible for creating Berserk). Setting aside how outlandish that sounds for a second, the fact we’re talking about an AI sourcing images without permission from an artist after his death and directly using them as building blocks in order to create something in his particular art style is genuinely pretty disturbing. Ultimately, that’s the crux of why AI art generation is so passionately despised; it can only create convincing looking, visually appealing images because it’s leeching off the creativity of real people without their consent. In all honesty, I don’t think people would have nearly as much of an issue with AI generated artwork as they do assuming that the knowledge base those algorithms were trained upon was sourced ethically, but the idea in-and-of itself of trying to “ethically” train an AI like this is more than likely untenable, especially considering how many different art styles the AI has to imitate in order to be effective. But honestly, it’s more than that. When you get right down to it, people don’t like the notion that something so traditionally personal and artistic as 2D animation was just spat out by a computer program and shoved into their faces. Part of why cartoons are impressive—regardless of whether or not they’re drawn in a western or eastern style—is because the viewer understands everything they’re seeing on screen was laboriously drawn and animated by hand. Already existing and common technology like 3D CGI blurs that line, but not to a degree where it’s upsetting for the average viewer because there’s still a different kind-of artistry happening there worthy of being respected, especially if that CGI is being used in the right context in order to enhance the whole of a particular animation. Supporters of AI generation software for things like voice acting, novelization and artwork argue that it’s about opening up avenues for creativity to a broader audience, lowering the barrier for entry for those looking to realize their creative vision in an audiovisual format. I, and many others, would argue that the development of software like this was never for the sake of trying to promote creativity or appeal to the average everyday user. It’s about streamlining and simplifying already existing processes. It’s about efficiency, speed and the elimination of existing roles for artists, animators, voice actors and writers; ultimately, it’s about giving corporations the power to do themselves what they would normally have to hire people for.

Now, do we necessarily have to be worried about an incoming dystopian future where all media is just AI generated by massive corporations without any need for human input, destroying creative professions utterly? Of course not—that’s just silly. But the underlying worry behind that thought—the idea that creativity and passion for things like writing, artwork and animation will die because it’s no longer considered a safe or viable career path—is a very valid one. As someone looking in from the outside of the industry, I think it’s easy to see people worried about whether or not they can carve out a future for themselves professionally as an artist or writer and say “Well, if you’re concerned as to whether or not your creative vision or your artistry will be respected, don’t try to turn it into a career.” But while there is some truth to that, I think the notion that any aspiring artists or creatives should expect to have their creativity compromised and its value disrespected when trying to pursue their passions professionally is a very dangerous one, especially in regards to the anime industry. Indie anime doesn’t really exist in the same way that indie games do; whereas most people familiar with today’s gaming landscape will tell you indie games are the way of the future and triple A titles have continuously regressed over the last decade or so, anime is an industry almost totally dominated by the giant animation studios at the top. It simply requires too much money, manpower and talent for any small, independent studio to make anything worthwhile in any reasonable amount of time. This is compounded by the simple fact that the vast majority of the anime that exists today is not original; the industry is almost entirely composed of light novel and manga adaptations, meaning that animation studios have to go out of their way to try and license those properties from their original creators. Put yourself into the shoes of a writer whose work has become a massive success story, prompting outreach from studios who want to animate and televise it; are you going to choose a small, unknown animation studio without the resources to make something you’re sure is going to impress, or a massive, renowned and well-respected studio who are more than capable of delivering something at a high-quality standard? Really good anime from very small studios is a rarity primarily for this reason; stuff like Happy Sugar Life is not anywhere near normal to see, and for as fantastic as Happy Sugar Life genuinely is, that show still failed and could’ve seen significant improvement visually! Of course, not every animation studio is the size of KyoAni or Studio Trigger, but there are amazing indie games out there like Stardew Valley created by only a single person; something like that is completely impossible for the anime industry.

If you are going to create anime (particularly as an artist or animator), you need to attach yourself to an already existing animation studio; it’s that simple. Even contract work will force you to collaborate with a studio. So the idea that your work and effort can just be co-opted by an AI specifically because the studio you’re working with is doing everything in their power to eliminate your role for the sake of trying to cut costs and push their projects out the door faster is deeply worrying. Worse, it’s not just that the studio behind The Dog & The Boy is trying to reduce workload and save money, it’s also that they’re making a public spectacle out of trying to glorify this technology. The reason why I can tell you the specific steps that went into their AI generation process is because they publicize themselves the fact it’s been used here, showing a visual representation of what exact steps and iterations the backgrounds present in the show went through and which parts were done by software versus which parts were done by hand, all under the horribly misguided pretense that the average viewer will find this technology “impressive”. No, it’s not impressive, it’s dystopian. It’s a god-damn slap in the face to the artists and animators who worked on this particular movie that their contribution is glossed over, whereas the contribution of a fucking computer algorithm is given the spotlight. We are quite literally witnessing a corporation try to condition and manipulate viewers into believing that this is acceptable so they can cycle out artists and animators who they would rather not have to pay, throwing their audience into a pot of cold water and making a sorry attempt to slowly but surely boil them alive. This kind-of shit is going to inevitably drive people away from the anime industry, which is already a career path traveled only by those who have immense passion for the medium when talking about it purely on the ground level. It’s upsetting for employees, it’s upsetting for viewers who want the assurance that what they’re watching was built entirely by human hands; in the end, the only people who aren’t upset are the investors and the guys at the top worried about little more than their bottom line, totally unconcerned with the quality of the product they’re actually putting out. From their point of view, choosing not to use AI generation software for moral or quality reasons would be a pointless, ridiculous act of pure charity; if it saves them money and lets them finish their product faster, why not use it? Why worry about the human element of a show being compromised by AI if that AI has to be assisted by humans? All the while, conveniently ignoring the fact that this so-called “AI” can only exist because of the work it’s stolen from thousands upon thousands of human artists in the first fucking place; if I’m being completely frank, there’s little to nothing truly intelligent about it.

A lot of what’s gone wrong with big companies in entertainment industries like video gaming or television has to do with the idea of what’s charitable and what isn’t from their point of view. Gaming is probably the most extreme example; sure, Elden Ring sold more copies than Call of Duty, but it still made FromSoftware significantly less capital because Elden Ring doesn’t have multiple monetization avenues. Many of the triple A video games of today are less complete games and more platforms for additional purchases; loot boxes, gacha systems, microtransactions and battle passes just to name a few. If it makes them more money to add those systems—regardless of whether or not those systems compromise the design integrity and fun of the game itself—why not add them? Surely it would be stupid to just gloss over such a genius money-making strategy—after all, games are a business, right? Triple A companies like Electronic Arts and Ubisoft aren’t doing this for charity! When you think about it from the company’s point of view, it makes a disturbing amount of sense, and it fucking sucks because it makes for an objectively worse and more predatory experience for we as the players. The same thing can be applied here to The Dog & The Boy (albeit to a lesser degree); there’s an underlying feeling here that even if the end product is acceptable, it’s difficult to enjoy it while knowing what the company is trying to do and how they’re trying to throw their employees under the bus. Do you honestly think any of what makes anime enjoyable for audiences is as a result of the big wigs at the top crunching numbers and bellowing marching orders down to their staff? Of course not; anime is enjoyable because of the actual artists, animators, writers, directors, musicians and genuinely passionate people who give a shit about making something worth your time. And here, those very same people are being given the middle finger and told their contribution is lackluster when compared to that of a fucking artificial intelligence. As someone looking in on the show and the controversy surrounding it from the outside, how am I supposed to support that? If anything, it worries me; there’s a constant fear in the back of my mind that now I have to worry about the integrity of every show I watch potentially being compromised. I’ve been deep in the trenches of the modern video game industry for over a fucking decade now, and I can tell you the last thing I want is to watch the anime industry go down the same dark path; the death of integrity, the advent of predatory money-making strategies and a total lack of concern for the livelihoods of employees genuinely passionate about making something great. I think I find this kind-of thing especially sickening because my integrity is my fucking life; I’m a critic for god’s sake. I grew up watching TotalBiscuit (still one of my greatest role-models today) and even in my day job I’m always concerned about approaching things from a “consumer-first” mindset, I actually had a moment just yesterday where I was walking back to my car after making a delivery, noticed the customer accidentally gave me a 50 dollar bill instead of a 20 and immediately rushed back to the door to confirm if it was intentional or not (which it wasn’t). Short term profit means absolutely nothing to me when contrasted against respect from the consumer for a reputation of honesty. That kind-of honesty and respect for the viewer or player is the reason why companies like FromSoft can succeed, because there are millions of guys like me who see an announcement for the next FromSoft title and start frothing at the mouth in anticipation, knowing that it’s going to be fucking awesome and isn’t going to try and pull any predatory bullshit. In the anime industry, I’d say Studio Trigger is a close equivalent; I am still beyond excited for their adaptation of Dungeon Meshi, that shit is gonna slap. If you’re curious why I praise them in particular, it was actually reported not long after the release of Cyberpunk Edgerunners that CD Projekt Red tried to censor some aspects of the show and Studio Trigger basically told them “fuck off, our way or the highway”, I have a lot of respect for that kind-of shit.

Getting back to the original point of all this, do I think AI generation is the devil and should be eradicated entirely? No, that’s probably a little bit too extreme of a viewpoint. But as things currently stand, there are too many issues with trying to use it as it exists today; it’s plagiarizing the work of artists without consent, there’s a false narrative being spun about how sophisticated it is and how artists will lose their jobs over it despite the fact it pretty much requires a babysitter to produce anything worthwhile at which point you might as well just have an artist draw something unique instead of struggling against an unfamiliar and deeply flawed workflow, the idea of using AI generation for an expressive, deeply personal entertainment medium like anime in-and-of itself is kind-of disturbing and easily capable of compromising the integral human elements of a show, and companies trying to glorify the advent of artificial intelligence while devaluing the contribution and talent of real artists will inevitably drive people away from the industry. As things currently stand this problem is still in its infancy, so we’ll have to wait and see how things continue to develop over time; to be honest, The Dog & The Boy in-and-of itself would be completely irrelevant and wouldn’t have even been on my radar if not for the amount of controversy surrounding it. Other examples of controversial AI generated media exist elsewhere; games with AI generated assets, AI generation software like ElevenLabs capable of imitating existing voices (which is actually at the point where it’s a bit scarily convincing), books like Alice and Sparkle that are almost entirely AI generated, etc. In all honesty, I find AI generated novella far more insulting than AI generated artwork; if you’re not writing something yourself, what’s the point in making a fucking novel then? I get the appeal for stuff like AI Dungeon—giving an AI ridiculous prompts purely to see what kind-of nonsense it vomits out and having a good laugh about it—but actually publishing a book written almost entirely by a computer program is just ridiculous, have some respect for the medium! We joke about a person having to babysit their braindead AI software, but a case like that of Alice and Sparkle is almost the exact opposite, a human being babysat by a fucking AI; and you know, I don’t blame the writer for his lack of creativity, but I certainly blame him for trying to use an AI to leech off the creativity of other people. And as far as ElevenLabs is concerned, it remains to be seen how that technology can be taken and translated into a professional project, but the idea it could be used as a complete substitute for actual voice acting talent is deeply worrisome. (Oh, and if I’m being totally honest, The Dog & The Boy honestly feels like its script was generic enough to be AI generated too). Anyways, I’ve talked about this long enough, imma go relax and watch some more WTF Is videos, be on the lookout for the inevitable uprising of artificial intelligence ladies and gentlemen.

Leave a Reply

Your email address will not be published. Required fields are marked *