The Dark Side of AI Art in Gaming

I wrote a long piece on artificial intelligence not all that long ago. It focused a lot on AI writing and my dealings with it in the marketing field.

TL;DR from that article (which is worth reading in full), AI as a whole isn’t all good or all bad, but a lot of the most popular uses for it are bad-faith for a variety of reasons, and the whole space is ripe for opportunists and grifts. The use I do get out of AI tools is not of the kind that involves typing in a prompt to ChatGPT or Midjourney (or one of their competitors), but rather in large-scale data analysis that’s beyond the ability of humans and informs my creative efforts in analytical ways but doesn’t do the actual creative parts for me.

I’d like to revisit the topic, because even in this short time, my own views have matured on it.

This isn’t a redaction of the earlier article, though. I stand by it. Rather, it’s that I don’t think I went far enough in some respects.

Because I see a lot more harm than good occurring in the gaming space as a result of AI content, both written and visual art. This is a microcosm of what I see on a larger level and the societal harms of its proliferation.

The Airing of Grievances

Rather than a narrative format, below you’ll find more of a stream of consciousness of interrelated points. Broadly, I’ll start with high-level ideas and work toward specific gaming topics. I’ll attempt some synthesis toward the end.

• Technologies are not neutral. Some will claim otherwise, that it’s the ways they’re implemented that bring harm or help. But their very foundation speaks to financial, societal and political realities. The ways these particular tools have been built is to produce plausible opportunities from which excessive hype can be used to sell software to people while disenfranchising experts, artists, and shifting wealth. It’s a grift, with any legitimate value (which it does occasionally have) being accidental, not intentional.
• Despite the proliferation of misinformation, excessive hype that oversells capabilities, and straight-up scams, artificial intelligence – as in, software that can “learn” and adapt as new information is presented to it – does exist and can have valuable use.
• I see the most valuable potential uses – within medicine, sociological trend analysis, media recommendations, and others – as wholly removed from creating an end product.
• These things are software tools and information predictors based on large data sets. They don’t have sentience, they don’t know they exist, or that we do, nor do they understand what they create. “Fancy automation software” would be a more apt description, but we enjoy dramatic overstatement in our media.

Corporate Interests in Selling AI

• Multinational corporations have, at this point, trillions of dollars invested in AI.
• This creates a need to justify that investment, which is why we see AI cropping up everywhere.
• Stated differently, AI being crammed into every crevice of our modern digital landscape isn’t from a perceived need or legitimate attempt to solve issues for modern consumers, but because of nervous billionaires who need to see their investments pay off via use of the technologies they’ve funded.
• If your devices feel worse for this fact, you’re not alone. The corporations know this too; they simply don’t care.
• The theme of “because we can” or “because we need to for financial reasons” will recur. Astute readers will notice this has nothing to do with artistic intent or passion, or usefulness of the technology. Where usefulness exists – which it does at times – it’s incidental to the true purpose behind its rollout.

Artists and AI

• Art is already democratized, in that anyone can pick up a pencil or instrument (or a thousand other tools) and start creating. The barriers to true creation are minimal in modern society.
• AI doesn’t democratize art; it merely devalues it.
• Some niche uses for AI as a supplemental tool can be shown that legitimately enhance processes for existing creators. Art, writing, coding, and so on. This is where productivity is being enhanced without plagiarized material being the source, and where the creative intent and execution lies with the original creator.
• An example of this is a master coder writing banal code with AI, then adjusting it personally and writing the harder stuff from scratch that AI won’t have useful priors on. This is reasonable use for creative purposes that still keeps agency with the expert.
• However, we should not lie to ourselves. These instances are not the norm. Not even close.
• Countless artists have explicitly expressed their non-consent to being used in training data, and have been able to prove that these wishes were violated.
• Millions of lobbying dollars are being spent to keep regulatory bills away from Congress here in the US. The reason for this is beyond obvious: legislation has not caught up to this technology, and companies know that their days of wanton data scraping are numbered if a close look is given to their practices by any legislators who aren’t in the pockets of AI conglomerates.
• When data has been scraped from other AI companies, spats and lawsuits have followed from the AI companies themselves. This isn’t about freedom of information or tools, it’s about exploiting loopholes in the system and maintaining beneficial double standards for as long as possible. They don’t want to share their information; they simply want to steal anything they can and then keep it private.

Ethics & Energy

• The energy needs for AI have sobering implications for global energy usage.
• Talking points to the contrary are refuted by the stated intentions and actions of megacorporations behind the majority of this usage. Many are pursuing nuclear power plants to be able to fund their AI growth.
• Water usage around new AI data centers has threatened to displace local homeowners and wildlife. The companies maintain their innocence in this regard, but the evidence that these centers are environmental plagues is piling up at a rate no amount of denial can realistically cover.
• AI usage is already disrupting energy grid usage in areas, and threatens international energy infrastructure if left unchecked.
• As a result, many of these corporations have rescinded their energy consumption pledges from prior years, as it becomes increasingly obvious that they won’t meet or even come close to their stated pledges.
• It’s dubious that they ever intended to meet these pledges (it’s good for PR, though), but increasingly they’re not even pretending at environmentally conscious energy use, or carbon neutrality, or similar pledges that routinely pop up in advertising.
• Pointing to other uses of energy is a disingenuous “what-about-ism” that plagues disingenuous discourse of many varieties. If there are other problems with global energy use, by all means, let’s discuss them. But don’t use them as a way to hand-wave AI issues.

The full picture it paints is one of corporate greed and opportunistic evil.

Creative Intent

An AI has no intention when it creates. Even though the animus of a prompter guides it, the output is based on work that had other intent behind it; repurposed without knowledge, it’s stripped of this intentionality and is thus creatively bereft to me.

This isn’t just about principle to me, though it’s that too. This has direct implications for the quality of the media we consume. The writing, the art, the games, whatever.

Because AI art is worse art. It can pass for human writing or art…at least at times (many times it’s stupidly obvious and bad, or perhaps simply generic and forgettable). And without a fundamental shift in how the technology is commonly used or works on a basic level, simply refining existing software tools isn’t going to close that gap.

An analogy I heard that I liked was a comparison to Fast Fashion. This is the global fashion industry that has prioritized cheap, accessible clothing items. We can make more clothes, at a lower cost, than ever before as a result. But the clothing does not hold up as well. It’s not as unique or as durable. It’s simply more convenient, requiring less advanced skill in its creation.

Cheaper, to scale, more convenient, but slightly worse than expert output? Sounds familiar, yeah? But it doesn’t paint a picture of progress so much as it paints a picture of scale and of devaluing craftsmanship. And there’s still a usable output! Which is the other darkly seductive part of AI. But it’s not particularly intentional.

Which means any time it’s being used, it’s merely a shortcut to an end result. It’s not about the process or eliciting the most from a creative work. It’s simply output. It’s content for content’s sake, in its basest form. Which turns art into something more akin to an assembly line, and I hopefully don’t have to explain why that’s not a step in a positive direction for the artistic sensibilities of individuals, communities, and entire civilizations.

If we value art only as a monetized end product, the majority of its power is stripped from it. Those pushing hardest for the commoditization of art seem incapable of admitting or perceiving art as a process, a dialog, and a vital organ pumping lifeblood through the veins of a society to keep it questioning itself, challenging itself and growing into something more human.

Implications For Gaming and Creativity

You get why it happens at least, right? A designer with an LLC is trying to self-publish on a shoestring budget, or a larger company with employees and offices to pay for needs to maintain a level of revenue to avoid layoffs or closures.

Prices on manufacturing and fulfillment never go down year-over-year, and so they’re staring at their budgets. And, oh, what’s this line item here? Art budget? Hmm…what would it look like if we dropped that down a bit?

Or a board game designer gets a couple neat suggestions from ChatGPT and decides to give in and start automating art, game mechanics, game names, and more. Soon, they’re just pushing a mish-mash of AI ideas with the goal of publishing, rather than pursuing a specific passionate creative intent that, yes, is with the intent of being published, but is spawned from a different, more creative process that’s much more likely to produce something unique and challenging.

Or a fledgling Youtube creator is looking at far more polished competitors and their own meager subscriber count, knowing they have years of hard work to sniff relevance on an industry level. Why not spruce up their production with AI instead of hoping they can improve their video editing skills over the next 12 months?

And that creator never improves their skills; their videos stagnate and lack artistic cohesion, but they’re produced more quickly, look a little bit better, and provide short-term gains in audience.

Somewhere the next Spielberg is typing a prompt into Sora (OpenAI’s video creator) instead of heading outside with their camera to shoot their first film. You’ll never see their AI stuff, because it will look like everything else produced by AI. Their unique artistic voice will never manifest into anything worthwhile.

But they’ll get more subscribers in the short-term.

And on the publishing side, jobs are lost. Art gets worse. And the practice becomes just normalized enough by addicted consumers who are willing to justify unethical behavior and slightly worse products to themselves on the basis of owning a new toy that looks fun to them. And so it continues, despite the often passionate backlash to projects that use AI art. The market is still there, in high enough numbers that it’s viable for many companies.

Not that this stops publishers from trying to dodge the issue. One common tactic from the playbook includes not disclosing AI art until it’s called out in enough numbers that it forces a public statement.

Another increasingly frequent one is to include some disclaimer along the lines of “The Project includes art that is a combination of AI generated material and the work of professional artists.”

Except something is often missing there: artist name(s), a portfolio, perhaps progress shots to confirm that the AI was just used to refine existing work. Indeed, variations on this line have been used by some creators where it’s then been proven that they generated something wholecloth with AI, with no artist input whatsoever. The “artist” is a poorly-paid intern typing in prompts; they used the line above only to avoid criticism.

It’s a line intended to deflect the worst criticism; it’s often disingenuous or is simply false. It’s PR spin to try to get a product across the finish line. Which again turns art into something that’s simply about the monetized product, not the human purpose for its existence.

People need to make a living, so this isn’t a rant against commercializing art. Creators should get paid. But that’s sort of the point, right? It brings us back around to true creators being paid to bring their visions to life. We devalue this when we decide they aren’t necessary, and slowly water down the art we allow to be sold to us in any form.

Where There’s Value

I mentioned a couple times that the most insidious aspect of all of this is that there’s legitimate utility with these tools at times. I’ll give you some examples.

• I’ve heard cogent opinions from medical professionals saying that AI tools have helped them catch diseases earlier. They worry about long-term effects of relying on such tools, but can’t deny that it’s helped to save lives.
• New vaccines are being tested that are the product of expert direction mixed with AI analysis. I’ll point out that this still includes subject matter experts in these fields, but it seems plausible to suggest we’ll have vaccines on the market that wouldn’t have existed without AI.
• On a much smaller level, I’ve set up a series of integrations in my day job between AI and other software tools. It’s allowed me to much more quickly create and edit content calendars, analyze marketing data, and then create and automate audience lists and reports from that data that inform my work and the work of my team. All of which is on a scale that I wouldn’t be able to do manually, and which collectively saves our team hundreds of hours per year. Importantly, it isn’t creating an end-product, and is still being vetted by a human at each major step. But it’s speeding up our workflow considerably in some areas.

I could cite others, but they’d follow similar patterns to the uses above. However, none of this should be seen as a “foot in the door” justification for use of AI tools in every facet of our lives.

If anything, this article exists to draw a clear line in my mind between the degradation of art that I see in AI, vs. the more analytical automation tasks that I feel it is better suited for, and which doesn’t rankle my creative sensibilities (even if it’s still not devoid of potential issues). Automation software is nothing new; I’ve been using it professionally since about 2010. An extension of that is comparatively benign, but of course that’s not all AI tools do, nor is it the majority of what they do.

Even Ed Zitron, arguably the biggest AI skeptic of our time, has written: “I don’t think LLMs, as a tool, are a grift. There are use cases, though those use cases are miniscule compared to the egregious promises and extrapolations made by the majority of the media and the executive sect, and absolutely nothing about them warrants the amount of money invested in them.” (full article link at the bottom). It’s the shoehorned bile that he takes issue with, of which there is a ton.

Meanwhile, anyone I trust who is both an expert in their field and who doesn’t have a vested financial stake in AI being the solution to everything, has told me that they use it for some specific, niche uses that add value to their work, and ignore the vast majority of AI applications in other software tools. And this comports with my own experience.

Conclusions

I dunno, I’m just a little bit tired of it all. I want to be excited by technological advances; for years I found joy (and still do) in exploring innovations in software for graphic design, video creation, audio creation, writing, and more. It’s made me a better creator and enhanced my ability to bring artistic visions to life. A few niche uses like those I alluded to above actually follow suit within AI!

I would be nowhere as a creator if it wasn’t for advances in technology that have allowed me to unlock more potential. I saw myself as a tech champion, not some Luddite.

So I have created art in each of those areas, and others not listed. And I still do! I make music, and I do cartography, and these blogs, and occasionally still make Youtube videos. I design board games and RPG supplements. I dance. Democratized creation tools, educational resources, and hosting platforms have made this possible for me, and I adore them.

AI creation doesn’t do that, though. It sucks the purpose from the process and leaves a husk of that deeper artistic value. The people pushing it or justifying it seem to either be those who are profiting off of it, or those they’re selling things to who are too embedded in consumerist cycles to stand up for something better.

The good uses I see – and have – for AI are not a part of that cycle; they might be on the edges, enhancing legitimate creators’ ability to do my work. But it’s still theirs. Not stolen, scraped, or amalgamated from existing art into something homogenized and purposeless.

This is one opinion, and not an influential one. It’s also far from the only one on the topic. But it’s mine, and as I see how AI is being pushed and practiced, I’ve gone from a more passive, “there are two sides” position, to one that is more firmly opposed to how I see it being used in gaming and other similarly creative spaces.

This is still not a blanket condemnation of AI, though. It’s more the specific ways that it’s been taken over by dispassionate corporate interests and mindless consumerist intentions that I can’t stand. There is a world in which I am actually enthused by thoughtful AI usage in creative processes. To date, though, we are not in that world, and so I must assuage myself by speaking my mind in the hopes that it moves the meter even slightly toward a state of general AI usage that I’d be more excited about.

I leave you with a few resources that I’m a fan of that take similar stances, which have helped to solidify my thoughts on this matter. Enjoy!

Meeple Mountain: Is This Really What We Want? AI Art in Board Games – by Andrew Lynch

Revenge of the Business Idiot – Ed Zitron

AI as a Fascist Artifact – tante