With the continuous growth and development of AI Large Language Models (LLM), more and more AI tools are starting to infiltrate industries that seemingly have no relation to AI. The video game industry is not one of those. AI has always been -in a very different way, mind you- part of game development. Every week we hear from more and more studios that are taking an ‘AI first’ mentality or embracing AI created content in some form to “optimize” workflows. This time around, it’s one of the big boys, so maybe we should talk about it.

From the CEO of Level-5 studios admitting to use of generative AI in the development process and reveal teasers of their games, to now Epic adopting an “AI first” production model for its upcoming projects with Disney, AI culture seems to be spreading in the gaming industry thanks to administrative mandates, rather than by choice of the developers themselves.

Last month at the cusp of Tokyo Game Show, Level-5 showcased its upcoming Professor Layton games and the long awaited Yokai Watch sequel, but upon closer inspection fans found traces of what seemed AI generated artwork. This raised some red flags as the studio has always been seen as the ‘studio Ghibli of gaming’, creating games that might not have the latest most realistic graphics, but the level of attention to detail and carefully crafted artwork gained the hearts of many, myself included.

Seeing what seemed like clear blatant use of AI art to showcase the remake of a beloved game like Professor Layton and the Curious Village was one thing, but then receiving a half apology delivered through a video that clearly used AI, that cost the studio a lot of the good will and love it had gained from its community over the years.

Akihiro Hino, CEO of Level-5, posted on X to apologize and clarify the extent of AI used both for the presentation and the games themselves. And like any good apology, there was a “However…” right in the middle of it, just under his very clearly AI generated avatar. As you can imagine, many long time fans of the studio didn’t quite take to Hino’s clarification that AI was used merely to make the presentation “more spectacular” or that “the time saved through AI-driven efficiencies has allowed our creators to dedicate more of their efforts to the true essence of creativity”.

Whether you are a game developer or a player, it is clear that a lot of studio heads and CEOs are starting to join the current generative AI craze. And while this means a lot of jobs being cut -as seen with Microsoft’s layoffs- a lot of developers fear it might also mean having smaller overworked teams working with AI tools. A recent survey found that “61% of European game developers are concerned that generative AI will place them under pressure to work faster and produce more”. The question is, will the product of said pressure be better or worse?

Epic Games’ recent stumble over a field of rakes while announcing its collaborative project with Disney will be an “AI first” production paints an even clearer picture of what’s going on. Companies and studios are no longer necessarily looking into firing people, rather giving the 4,000+ developer teams AI tools to crank out new games and products as fast as possible.

Not very long ago, AI meant something very different in the gaming industry. Custom systems created to dictate how enemies react, maybe particle systems or pathfinding logic so your NPC companions don’t get stuck on a corner trying to follow you through a dungeon. Artificial intelligence, in the most basic sense of the word, has been a part of game development for as long as games have existed. Perhaps that is why a lot of industry veterans and studio heads are more open minded when considering newer AI tools. Although there is quite a big difference in the nature of these newer AI tools, compared to what has been used in the industry for many years.

Programming is at the heart of game development. For all the amazing art or brilliant puzzle design that a game can showcase, programming truly is the glue that holds it all together and makes it actually work. So, it is not hard to see how as soon as AI agents got slightly better at coding, they were introduced into the game development process. After all, programming can be one of the hardest and most time consuming parts of game development. From actual programming of mechanics, to testing and bug fixing, games spend the most time in the programmers hands.

But even with the most advanced AI programming tools out there, most of the industry programmers recognize that it is not at the level of ‘prompt and publish’. AI can assist with the base creation of code, a sort of jumping-off platform for the rest of the code. It may even help find bugs or syntax errors, but it most certainly cannot code a game without an actual human -with programming knowledge- going through it and adjusting where the AI is unable to fill the contextual gaps.

While it may not be said out loud, it is common practice for programmers in big studios to use AI tools to assist their job. But that is the key word: assist. Production pipelines ,like the one announced by Epic, that aim for “AI first” with “no hand-written code” sound like the perfect recipe for a messy game at best. That is without taking into consideration the simple fact that game development is an iterative process where programmers, designers and artists push each other forward through collaboration, creativity and improvisation. Things that AI is not famous for.

To me, the idea of AI assisting programming work in game development doesn’t seem ludicrous. It merely sounds like adding extra unnecessary steps that purposefully hinder the honing of human skills, while making it harder for future or less experienced developers to actually develop the creative and problem solving skills that allow them to catch AI’s errors in the first place. So for now, it might be useful, but in 5 or 10 years time, when programmers no longer have the know-how and actual programming skills that veteran developers have, we might start getting a lot more messy ‘vibe coding’.

This argument is even stronger for the art side of development. If artists and developers start using AI to make the art of their game “more spectacular” or “save time” making it, like Akihiro Hino suggests, there will eventually be a time when all games look the same. That is, with the same flaws and the same uninteresting shading and cartoon-like style. Even when artists use AI tools to assist the creation of assets, there is a human spark lost throughout the process. And the same idea of future artists losing the opportunity to grow and hone the skills that their predecessors worked hard to develop applies. The same predecessors that sadly got the fruit of said hard work used unlawfully to train the AI models that have now hindered future generations. Although, the unlawful training of LLM’s with millions of stolen art pieces is another topic. But, should it be?

I think it should absolutely be part of the conversation. It is precisely people like Akihiro Hino, CEO of a studio that was wronged by AI companions unlawfully taking the art that was lovingly and painstakingly created by the artists at Level-5, the ones that should be using their platforms and communities to push back, not falling in line with them.

One creative that has been very vocal of his opposition to AI in the creation of art and media is film-maker Guillermo del Toro. He claims that it’s not just because -and I’m quoting him- “AI has demonstrated it can do semi-compelling screen savers”, but because it robs the next generations of creators of the history of their medium. At a recent Comic-Con panel for the re-release of Pan’s labyrinth he said:

“It’s not about who gets the job, it’s about [we are] protecting a lineage of art. If we cut a generation of people from learning their craft, you are cutting the rest of the history of that medium away from them, for what?”

If we, as in the gaming industry as whole, choose to start using AI generative tools to create art or code our games, using AI as the main creative source, we are cutting future generations of game makers from actually learning how to make games and the history of gaming as a whole. The only way that those future developers will learn of the past of our medium is by half-baked remasters of games that are important to the history of gaming, but that won’t teach anything to the teams remaking them, or to the new players experiencing them for the first time.

As the AI push keeps on growing, and will probably try to enter more and more of our lives, we need to start pushing back in whatever way we can. Both as consumers and as developers who are part of the industry, and who want to see the legacy of gaming history be preserved, not fed into a machine that will reorganize permutations and feed us a semi-compelling screen saver.