By Jon Scarr
Jack Buser is Global Director for Games at Google Cloud. He wants the games industry to think back to the PlayStation 2 era. He isn't arguing that games need to look or play like they did 25 years ago. He believes studios once had more room to experiment.
In a new Google Cloud article, Buser argues that modern development has become expensive and time-consuming. He believes that pressure encourages publishers to rely on established franchises, sequels, and remakes. His answer is AI. He believes AI can reduce repetitive technical work, leaving developers more time to take creative risks again.
I like the question behind that argument. Games don't need another excuse to become more predictable. But making them cheaper or easier to produce doesn't automatically bring back the conditions that made the PlayStation 2 era so memorable.
There are a few much harder problems sitting underneath it.
AI Can Remove Work Without Replacing the Creative Part
The most convincing part of Buser's argument has little to do with AI-generated artwork or dialogue.
Capcom is already using AI agents to perform more than 30,000 hours of autonomous playtesting every month. Google's case study describes agents that inspect visuals, predict where technical problems are likely to appear, preserve internal development knowledge, and sort through complex data. The purpose is to take repetitive technical work away from developers so they can spend more time on creative work.
That distinction came up directly in the LinkedIn conversation around Buser's article.
Alex Van Winkle pointed to narrative, characters, plot threads, and choices as a common thread in games that stand the test of time. Buser agreed that intentional, human-designed constraints are what make those choices matter. His response was that AI should expand what storytellers can build around those choices rather than replace the person creating them.
That's a much more useful AI discussion than pretending every creative task needs automation.
If a studio can automate repetitive test runs without replacing a writer, artist, designer, or programmer, it's easy to see the benefit. The difficult part is making sure that line survives once cost savings become part of the conversation.
Living Games Could Change More Than Development
Buser's bigger idea goes beyond production. He calls the next step “living games.”
The phrase sounds like another industry label until you get into what he means. In his LinkedIn exchange with Aniket Sinha, Buser drew the line at persistent contextual memory and agency. An NPC wouldn't simply react to what you just did. It could remember an earlier interaction, carry that history into a later encounter, and change its relationship with you.
Google points to several early examples. Dragon Quest X Online has experimented with Chatty Slimey, an AI companion designed around an evolving relationship. YOU vs. Zombies generates personalized heroes through its Ritual technology. Ubisoft's Teammates prototype lets you speak naturally to AI squadmates instead of selecting commands from a traditional menu.
There's potential here. A character remembering how you treated them five hours earlier could create interactions that aren't possible when every branch has to be manually mapped out.
It also raises the question Sinha put to Buser. Haven't games been adapting to us for years?
Dynamic difficulty, procedural generation, simulation, and reactive NPC behaviour aren't new. Buser's answer is that persistence is the dividing line. That is a useful distinction, although we're still talking about an idea that's being explored rather than an established new category of game.
More Games Won't Fix Discovery
Richard Bates raised what may be the biggest problem with the whole PS2 comparison.
If AI lowers the cost of making games, more people can make them. That sounds great until you remember how difficult discovery already is.
I've covered this problem before on 4ScarrsGaming. Faster production means more games competing for the same amount of attention. Making another interesting game is only half the job if nobody can find it.
Bates made essentially that point in his LinkedIn reply. Lowering the barriers to creation doesn't automatically lower the barrier to finding an audience. Buser agreed and said connecting games with the right audience is another major issue the industry still has to work through.
That makes the idea of recreating the PlayStation 2 era much harder than automating development tasks.
The PS2 didn't exist in today's environment. Games now compete across digital storefronts, subscription catalogues, mobile platforms, social media, and livestreams for your attention. Even if AI lets a small studio build something strange and original, that game still needs a way to reach you.
The Industry Isn't Sold on Generative AI
Research conducted by The Harris Poll on behalf of Google Cloud is extremely optimistic. Among 615 game developers surveyed in five countries during 2025, 90% said they were already using some form of AI in development. The same research found 95% believed AI was reducing repetitive work.
That isn't the same picture you get from GDC's 2026 State of the Game Industry survey.
GDC surveyed more than 2,300 industry professionals across a wider range of roles. More than half, 52%, said generative AI was having a negative impact on the industry. Negative sentiment was particularly high among people working in visual and technical art, narrative and game design, and programming.
Those two surveys aren't measuring exactly the same thing, so putting the percentages against each other as if one disproves the other would be misleading. One thing they do establish together is that AI use and enthusiasm about AI aren't the same thing.
Someone can use an AI tool to summarize technical information and still strongly oppose replacing creative work with generated material.
The Google Cloud study reflects that tension too. It found concerns around data ownership, privacy, AI-generated content ownership, and the cost of integrating the technology.
Creativity Is Still the Part Technology Can't Guarantee
I find Buser's PlayStation 2 comparison more convincing as a goal than a prediction.
If AI can remove tedious development work, improve testing, and help smaller studios compete with established companies, that's useful. If it eventually lets designers create characters and worlds that respond in genuinely new ways, even better.
None of that guarantees another Katamari Damacy, Shadow of the Colossus, Guitar Hero, God Hand, or whatever strange game someone hasn't thought of yet.
Technology can lower the cost of experimentation. It can't guarantee someone takes the risk.
And even when they do, discovery is waiting on the other side.
That is the part of the PS2 era I wouldn't want the industry to misunderstand. Creative freedom wasn't valuable because studios could simply produce more games. It mattered because talented people were able to take strange ideas, deliberately shape them into something with an identity, and get those games in front of an audience.
If AI can give developers more room to do that without becoming the creative voice itself, then Buser's golden-age argument becomes much more interesting. If it only increases output, we're going to have a lot more games fighting to be noticed.
Comments
Post a Comment
Comments are moderated. Please stay respectful and on topic.