I get a lot of AI games pitched to me, so I have had to carefully consider what AI-made games mean to me, and whether or not I should cover them here on my site.
As someone operating in an industry that has been absolutely destroyed by AI, and who struggles to make ends meet directly because of the technology, I have a somewhat biased starting point to contend with, but I want to be open about that, too. I acknowledge this is a nuanced argument, and not always as crystal clear as either side would have you think.
Having said that, watching as a huge number of my industry peers lost their jobs, and seeing whole sites go down, has been painful. I’ve even had to watch recently as some anonymous clickfarm bought out Gamereactor, a site I worked at for a decade and that contains a huge portion of my work.
Gamereactor has now been repurposed into a backlink farm for gambling and casino sites. The staff were all sacked, and the new copy appears to be written by AI and then aggregated across the various branches of the site. It’s a disgrace, but it’s also another sign of the times, a moment we all had just enough time to glimpse before we lurched forward into the next debacle.
Another of my previous employers, Network N, has been shedding staff and making fresh redundancies recently. I’m under no illusions; I would have almost certainly been among those who lost their jobs if I hadn’t been forced to leave by my health. My audience development and content strategy roles were “nice to have” additions to a company in a growth phase, but as belts tightened post-Covid, and then prompters using LLMs came in and ate up every single piece of low-hanging fruit out on the free Internet, the old media landscape that Network N had previously thrived in dried into a barren desert.
It won’t surprise you to learn that my experience with AI has informed my decision not to use AI-powered copywriting in any capacity on Rogueliker. I take pride in writing every word myself. It’s important to me personally that this site bears my voice. That simply wouldn’t happen as I wanted if I outsourced half of the work to AI.
Don’t get me wrong, sometimes it’s a tempting proposition. I could easily train the AI on my own site and voice, ensure it had a passable likeness to my writing, and then use it to keep my articles up to date, and so on. There are definitely genuine benefits to the technology, and if you like AI, I’m not here to yuck your yum. I’ve stayed strong and done things the old-fashioned way, but I acknowledge that is a personal decision.
Which brings me to my point, and another tangent.
Back in 2012, when I was still learning my craft as a games critic and news writer, a game came out called Gettysburg: Armored Warfare. It stuck in my memory because, at the time, the publisher (Paradox Interactive) made a small fuss out of the fact that it was the work of a lone developer.
If memory serves, it wasn’t a very good game, and I have no intention of redownloading it to confirm that. I do, however, remember that it was highly notable that one person had managed to make such an ambitious title alone. How could one person do all of those jobs to a high enough standard for their game to compete? It turns out, in that case, they couldn’t.
Yet, since then, we have had plenty of solo devs proving that as long as you scope out a project properly, a one-person developer can make a special game. Balatro remains an all-time favourite of mine, and the game itself is the work of one person, with the caveat that the soundtrack was outsourced and an experienced publisher brought it to market, and then loads of people playtested it. The point is, it’s possible to make a solo game, but great solo developers like LocalThunk and Lucas Pope are the exception, not the rule.
These days, however, every third game pitched to me is a solo-developed game. If we were getting a constant stream of original titles like Balatro, I don’t think anyone would really mind if I wrote about them, but that’s not what is happening here, in my opinion.
The reason behind this massive influx of solo-developed games is AI-powered tools, which have simplified elements of the process to the point where nearly anyone can make a game. In theory, at least, I think this is a hugely exciting development, and I appreciate the democratisation of creativity that the AI boom has created, even if I have personally paid a price.
If you want to make a game using AI, I think you absolutely should. I think doing so would teach you so much about the art of making a game and would put you in a good position to explore a career in game development. I don’t think you should be trying to sell that game, given the obscured nature of how some of the code/art/music was created, but I think AI-made games absolutely have a right to exist and be played.
The problem with making a game using AI tools is that you lose so much of the human element in the process of rapid development. A solo developer doesn’t have a fellow creative or team to bounce ideas off, to challenge poor decisions, and to push the developer into making things better – instead, they have a digital sycophant confirming their ideas and making them happen almost instantly.
By fast-forwarding through the development of a game with AI, you miss out on the friction that is created by the passing of time. Not only that, but you also miss out on the sharing of ideas with other people around you. Some ideas need to mature to be at their best. Some concepts need to be cut before they’re too entrenched in a grander design. The bad ideas that you have along the way are often what determines what your game is not, and this is just as important as what your game actually is.
I’m nearly always initially sceptical whenever I see the words “solo developer” in a pitch these days, because all too often, this means that the game has been made using generative AI.
This is a red flag for me, first because it tells me that there is a small chance that the core ideas or even whole parts of the project have been completely lifted from another game.
Alternatively, what is more likely, is that this is the work of a developer who has decided to hit fast-forward on creating certain aspects of their game. In making that decision and using AI-powered tools to achieve these results, your game might look the part, but there’s a good chance it’ll be missing a certain texture that can only be achieved through the process of sharing ideas with other people.
Therefore, not covering games made with AI is a stance that allows me to apply a level of quality control over the games I spend my limited time looking at. By simply saying “no”, I can instead focus on games made by people that I know have been thoroughly tested.
The other reason I skip AI-made games is ethical, but it’s aligned with this point, too. In my opinion, having a development team made up of coders, designers, writers, artists, and musicians allows for the cross-pollination of ideas that is missing from games that are made by one person and an entourage of AI-powered tools.
My final ethical point is the one I made in my headline. A development team made up of several highly trained creatives has earned my attention through their collective craft and hard work. They will always be my priority when deciding who to champion.
Not only that, but as I’ve learned personally, it’s truly horrible watching AI gut your industry and steal food from the table everyone eats from. I don’t want to enable that in a professional world I adore and respect.
If I have to choose between a game made by a solo dev with AI tools and one made by a team of people, or even someone working alone on a project for years, learning how to wear many hats along the way, I’m never going to choose the game that was made with AI because of the human cost of that decision.
Not only is there a small but real chance that the AI game was built on the backs of the games it’s competing directly against, but there’s an even better chance that the AI game just hasn’t had the rigorous testing it needs during the development process.
Furthermore, it’s simpler and ethically clearer for me to extend my policy to include major studios who have adopted the technology. Why cover things made by AI when there are so many great games made by humans that hardly anybody knows about? This flood of AI-developed games has made discovery harder than ever before, as the ongoing Next Fest has proven quite unintentionally.
In this life, we can have at most two out of three things, but never all three. You can have something fast, you can have it cheap, or you can have it good. By focusing entirely on the first two things, in my opinion, AI-made games will never be as good as human-made ones, which will, by their nature, never be quick to make, because, as Epic Games wrote just last week, programming is hard!
