Industry discussions about integrating AI into development processes appear to have quietly but firmly entered a new phase.
The question is no longer whether this technology will be used, but where the line will be drawn. AI tools exist and will not cease to exist; we need to figure out what uses are appropriate and acceptable to consumers.
At least that's what I've been told repeatedly over the past few weeks, and frankly, I'm starting to wonder if this apparent fait accompli is actually just a strategic line of communication pushed by one of the expensive PR firms that are doing very well in the crusade to improve the public image of AI.
It's a smart line because there's some truth to it: these tools are real, they're useful in some situations, and the technology won't become “uninvented” now that it exists.
However, it is based on the suppression of a very important difference. “AI” is an all-encompassing term used here to refer to virtually any computer system that is based on a learning process; not only a wide range of different use cases for large language models (LLMs), agents and other generative tools like Transformer, but also all sorts of much more established and proven use cases for other, older machine learning algorithms.
So yeah, it's a pretty clever rhetorical trick. Smart image scaling algorithms? AI. Voice recognition? AI. Code completion in your integrated development environment (IDE)? AI. Spell checkers, voice assistants, smart lasso tools in graphic editors? AI, AI, AI.
“AI tools will not cease to exist; we need to figure out what uses are appropriate and acceptable to consumers.”
The argument is how one can be “against” AI; What kind of Luddite do you have to be to deny developers the ability to use all these tools?
I've even heard people – who really know better – try out the outlandish claim that the gaming industry has always used AI because enemies and NPCs in games have had “AI” since the dawn of the medium, as if the little imp from Doom could be found throwing his low-pixel fireballs somewhere high up in the ChatGPT family tree.
The problem is that while this rhetoric may muddy the waters significantly in terms of internal debate or online discussion, consumers actually seem to have a pretty clear understanding of what they like and don't like from an AI perspective.
Nobody really cares whether programmers include LLM-based autocompletion in their IDEs. Nobody is upset about your deep learning scaling algorithm. What they care about is, to use the modern word, slop.
AI discards—assets, be it art or audio, created using model cues rather than created by humans—are a major concern for many consumers.
Of course, there are different levels of sensitivity to this. Some consumers hate it for environmental reasons or for moral and ethical reasons, since we haven't come to any solution to the whole “these tools were created by the largest act of brazen intellectual property theft in human history” problem.
Others just don't like the way it looks. AI garbage is often quite recognizable to at least some portion of consumers. This is regularly highlighted on social media by the deliberate uploading of low-resolution images and videos to mask the overly glossy sheen of the AI. Games don't really allow for this kind of obfuscation; AI-generated assets can usually be seen in all their glory in high definition.
For other consumers, of course, this is not a problem. We recently reported on How uncontroversial were AI assets in the mobile gaming marketat least until now. This is no doubt partly due to the different demographics of mobile game players, but it also depends on the medium: after all, people are much more used to seeing AI-generated content on their phone screens than in PC and console games.
Oh, and there's one more small point: almost all mobile games are free to play – because when trying to figure out where consumers draw the line and what they consider acceptable, one takeaway from the last week or so has been that a lot of consumers do draw the line firmly at paying full price for premium games with crappy AI resources.
Call of Duty: Black Ops 7 is facing some pretty harsh consumer backlash; this is not only because of the AI assets, but also use of very explicit AI resources in key parts of the game became a rallying point for consumers disappointed with the last game in the series.
Manage your cookie settings
It's not just that the assets themselves are bad – although, to be clear, they are terrible, right down to using that terrible Studio Ghibli-type filter that for a few weeks last spring was the most popular way to tell the entire Internet that you have the media literacy of a drunken seagull.
Moreover, consumers seem to be irritated by the very attitude that these assets involve.
That an expensive game, part of a franchise that has made billions, has such a lack of respect for players that no one sees a problem with using AI instead of human artists for key assets, including some of the banner art that players earn for achievements. Feed them slop; they won't notice the difference.
Of course, some people won't notice the difference, while others won't care. But there is a very significant portion of consumers who care, and this attitude is not simply driven by some left-wing ethical position on artificial intelligence or data centers – it is something much more fundamental and is likely to be a problem for any segment of the industry hell-bent on adopting this technology.
Consumers crave authenticity. They value “realness” and are willing to pay for it—whether it be food, clothing, experiences, or anything else, including media—this is an absolutely fundamental truth about how consumers think.
People pay more for authentic branded products than for copies; they despise fakes. “Handmade” commands a higher price than the most precise and efficient machine in the world. We pay more for local crafts, for art and music that makes us feel connected to the creator, for food that feels authentic and “real,” connected to place and history.
This is an instinct that permeates all areas of consumer activity. From the ridiculous prices of a brand like Hermes to the handmade souvenirs you overpaid for on holiday because they had a story behind them; we crave authenticity. For most of us, there is nothing more valuable than being “real,” and nothing worse than being “fake.”
“There is no faster way to reduce the value of your brand than to try to ensure authenticity and offer something machine-made.”
Of course, this doesn't mean that there isn't a place in the market for fast food or fast fashion, where we turn off those instincts and just go for something cheap that fills the gap. (With the greatest respect to the mobile sector, developments in the last decade have completely relegated it to the category of “fast games”.)
However, if you want to sell premium products and convince people to pay a premium price for them, you don't have to take such blatant shortcuts.
There is no faster way to reduce the value of your brand than to try to provide authenticity or reality and instead offer something artificial and machine-made. Consumers feel it instantly; it feels cheap and, worse, disrespectful.
The invention of generative artificial intelligence has not changed the fundamental relationship between authenticity and reality and consumer perception of premium products. It also didn't change the most basic trade-off that everyone in any line of business must face time and time again; “cheap, fast and good; choose two.” There's always a catch to anything that offers all three, and AI is no exception.
Companies that want to implement AI into their development processes (I'm ignoring the blathering rhetoric about IDE code additions and image scaling, I'm talking about building assets with generative AI) need to sit down and seriously ask themselves what kind of business they think they're in.
If they can honestly answer that they provide a quick and cheap experience for consumers who have no interest in how the sausage is made; if they actually accepted comparisons to fast food and fast fashion; then of course there may be a case for AI.
On the other hand, if you believe that your company provides premium products (and if you price them at $70, I would advise you to believe that by default), then you need to evaluate the risk you are taking.
“People pay for authenticity and will hate you for trying to dupe them with slop.”
Like a luxury clothing brand that advertises as “hand-sewn” but actually rolls it off the factory production line, you are one leak, one consumer backlash away from completely losing your brand's premium personality and your ability to maintain your prices.
This is where the line is drawn, and all the PR campaigns and clever rhetoric in favor of AI will not change it.
People pay for authenticity and will hate you for trying to dupe them with slop. This is a pattern of consumer behavior that pre-dates artificial intelligence, and it will take over several billion LLM tokens and a ton of investment to change such a basic part of human nature.
