It appears to me that big game studios tried and failed to do the same for video games. The "quadruple A" Ubisoft title Skull & Bones had a reported budget of around $200 million, while other sources reported that if you include the cost of various restarts the actual budget was much higher. But Skull & Bones didn't sell very well. Meanwhile games like Clair Obscur: Expedition 33 had a budget of below $50 million, and did a lot better in sales. Hollow Knight: Silksong had an even lower budget and even better sales. And they didn't have to pay anything for marketing, the game just sold by word of mouth.

The difference is in the middle men in these businesses: Cinemas can only show so many films every week, so they have to select those that are likely to draw the biggest crowds. Steam doesn't have the same limitation in the number of games it can release every week, so it just releases everything and lets the customers choose freely. Given free choice, the customers don't necessarily value big budget games higher than low budget games. Which makes the AAA video game industry look silly sometimes, as they lose money on big budget productions, while other studios rake in big profits on much cheaper to make titles. As soon as you remove the gatekeeping effect of big budget productions, big budgets turn out to be just big risks.

Now if you look at the last 50 years of video games, from Pong to Death Stranding 2, budgets for video games have certainly gone up a lot. But if you look at the details on what exactly those budgets were spent for, the lion's share obviously went into artwork. If you compare Civilization I with Civilization VII, it is very obvious that graphics have evolved a lot over time, while the improvements in gameplay are less remarkable. As a player of strategy games, it is one of my pet peeves how little of the development cost goes into the specific artificial intelligence used to run the computer opponents. The goal wouldn't be to make the computer unbeatable, as that wouldn't be fun, but certainly more could be done to make the AI a bit more varied in responses, and less vulnerable to cheesy exploits.

A very different sort of AI is more likely to become more prevalent in video game development: Generative AI. Big movie companies like Disney are currently at the forefront of legal battles against AI. And in game development it is the people employed by game companies that are still doing a rather good job of convincing the public that they don't want AI images in their games. That isn't going to hold forever. And while some of their arguments of theft are certainly justified, the fight of big movie and big game companies against generative AI is mostly motivated by self-interest. If most of the budget goes into artwork, it is artists that have the most to lose. AAA video game studios are still trying to gatekeep game development with big budgets, and generative AI would enable smaller studios to achieve much better looking games for much lower budgets. The big video game companies can't afford customers to judge games by their gameplay merits, because that would just financially sink them.

Steam doesn't have a ban on AI graphics in games, it only requires developers to disclose the use of AI graphics. That seems pretty fair to me. While I see the issues of deepfakes and theft by generative AI producing art "in the style of ...", I don't believe that every image generated by AI is automatically theft. If a game developer asks AI to draw a generic orc, the resulting orc image is certainly somehow related to all other images of orcs out there; but if the game developer would ask an artist to draw a generic orc, the resulting orc image would also be somehow related to many of the same existing images. As long as you can't draw a straight line from an AI image to a specific artwork or artist style, I don't see this as a problem. It should be up to the customers to decide whether they like the resulting images, whether they are "good enough" for the purpose of gameplay, or whether generative AI "lack of soul" or uncanny valley effects result into something people don't like.

So generative AI, with some supervision to assure no specific artist is ripped off, to me is a chance to make equally good video games at lower budgets. Especially games that don't just sell on artistic appeal, but on gameplay. We don't need to live in a world where half of the indie games are using pixel art. And I am sure that generative AI could also be used to turn interesting art from human artists into video game animations at much lower cost than today. I would rather see games competing on their merits on platforms like Steam without the gatekeeping of big budget artwork.