Ready player infinity: How AI is bringing living games and real-time content to life

For all the blockbuster games and indie hits released each year, more than half of all playtime is devoted to games released six or more years ago. This makes breaking through with new titles an uphill battle for many game companies. Wresting the attention of players still immersed in their beloved classics requires something radical.

Against this backdrop, AI has emerged as a transformative force, promising to reshape not just how games are made, but what games can be. In fact, a new survey, conducted by The Harris Poll on behalf of Google Cloud, confirms this, with a staggering 97% of developers surveyed believing that generative AI is reshaping the industry, from ideation to programming to, yes, gameplay.

To learn more about these seismic shifts, we sat down with Jack Buser, who leads the games unit of Google Cloud’s strategic industries team. Buser has been involved with games fore more than three decades, giving him keen insights into how AI is poised to revolutionize game development and the very nature of games themselves — a trend he's identified as "living games."

You've experienced several major industry shifts firsthand over the past three decades. Many understand the potential for AI to accelerate game development. You believe we're on the verge of something much bigger?

Jack Buser: If you look at the history of video games over the past few decades, it's primarily been a story of evolution in gameplay and graphics. It’s often very familiar IPs and genres getting released with better and better graphics, with occasional gameplay innovations that evolve existing genres. But every once in a while, a technology comes along that fundamentally redefines what a game can be.

This doesn’t happen with every formative technology. Remember FMV games? Stereoscopic 3D? You probably don’t, because plenty of technologies like these have come and gone, but all signals point to AI radically transforming the types of experiences a game designer can create and how a player interacts with them. The Harris Poll data backs this up, too: The survey found that 94% of developers see gen AI as a positive force for innovation, and 90% are already using some form of AI in their workflows.

In the early days of AI for games, much of the energy was around getting tools into the hands of development teams so they could work faster and more efficiently. Google has great solutions for that.

But what’s really going to change the world more than any of us realize is this idea of using AI during gameplay. We're going to see entirely new genres pop up. Disruptors we've never heard of before will create experiences that break through. This may dramatically shift the makeup of the entire industry. And if you're not thinking about that now as a game designer or a game team, player expectations may shift more quickly than you can adapt.

Live service games were revolutionary for the industry. Why is AI — unlike other technologies that have come and gone — poised to drive the next fundamental shift?

Buser: For decades, the games industry was effectively a consumer packaged goods industry. A development team would create a game, make thousands or even millions of physical copies on little cartridges or discs, slip those into a plastic clamshell case, and distribute them to brick-and mortar-retailers around the world. Once players checked out at the register, their relationship with the developer ended.

Around the 2010s, games started to become platforms. Developers learned to continually update their games with new content to keep players in it, selling in-game items, in-game currency, new levels, add-on packs, and expansions. This started largely in mobile games, and moved to consoles and PCs over time. Eventually, continual updates became fundamental to the game design itself, regardless of platform. And now the biggest games in the world never turn off. People play day after day, month after month, year after year.

When you're operating a live game, your development team is watching what the players are doing in the game and figuring out what they like using sophisticated tools — in fact, many of them are using Google Cloud for analytics. They then create new content based on what they've learned about those players. That cycle of a player’s behavior, analysis of that behavior, and finally creating and releasing new content back into the game based on that analysis, can take weeks or even months.

Enter AI. AI promises to do that analysis and content generation much more quickly than only a human could. If a developer has AI running during the game, and it’s helping figuring out what the players want and then creating the content, they can turn that loop around almost instantaneously. Developers still have traditional live operations, but complemented by this much, much tighter AI-driven loop, which responds to player behavior in close to real time.

What does this look like from a player's perspective?

Buser: It can take the form of characters in the game that can talk to you like we're talking right now. It can take the form of a game that's adjusting its design and content in real time based on what it knows that you — or players like you — want. It can take the form of assistants that, realizing you're stuck in the game, seamlessly help you through a difficult moment while keeping you fully immersed in the game world.

Many early examples of real-time AI inference are already in production, including some beloved and notorious characters that can converse with you on the fly. In other words, they don’t follow a preplanned script with limited dialogue. It's a really, really cool experience for players. But this is just the tip of the iceberg. In time, AI experiences won't only be added into existing games. Designers will weave AI into the fabric of the game design itself.

How should companies navigate this transformation?

Buser: I've spent my whole career in the games industry, I love it so much — but it’s not everything it seems from the outside. What looks like this bleeding-edge innovation machine, it can actually be quite conservative, surprisingly so. The industry often struggles with pivots to new modalities, new forms of entertainment, new technologies. We do well with iteration, but transformation can be very, very challenging.

The shift from cartridge to CD-ROM is a perfect example. You had so many developers going, "Hey, what do we need 600 megabytes of storage for? That's ridiculous." Maybe they'd put an orchestral soundtrack in their game to fill up the disc, but they weren't going to risk changing their game design. Boy, were they wrong.

Another example was the shift from 2D to 3D graphics. A lot of the game developers who said, "I'm not so sure about this 3D thing. Maybe I'll just take my same old 2D game design and try to create a 3D version of it" aren't around anymore.

Now hearing this, you may be saying to yourself, "Oh my gosh, we are wholly unprepared for this." We should all be saying this because we're all wholly unprepared. We've never done this shift to AI before — but that doesn’t mean we can’t. In fact, many of us have tackled similar technological transitions, and even built our success on them.

To do so, we have to start building proofs of concept right now. We have to understand how our development pipelines change and how players react to different types of AI-driven experiences. All these things are a learning process for our industry. Some companies are leaning in. If you're not, it could pass you by. You could find yourself building a video game that's antiquated within four or five years — with no ability to pivot fast enough to catch up.

Many game studios have concerns about IP protection and trust. How would you reassure them?

Buser: In the early days of AI and games — two years ago! — the number one challenge we had was this: game companies were terrified of stepping on somebody else's IP accidentally, or, in using these models, of leaking their IP back out to train the master model.

When game studios messed around with consumer-grade AI tools not built for professional environments, their legal departments said, “Stop that immediately!” Many studios even banned generative AI until they had it figured out.

This is where working with enterprise-grade AI is a necessity — and what Google has been building since day one. When we bring AI to game developers through Vertex AI, it’s enterprise-ready. First and foremost, if a studio uses a generally available version of a foundation model, like. Gemini 2.5 Flash or Pro, and somebody comes after them with an IP claim, we indemnify them. And Vertex AI doesn't use your Gemini prompts or its responses as data to train the models.

I remind game companies that Vertex AI is not a game development tool. It's a tool used across all industries, including massive financial institutions, healthcare providers, and government agencies. It's more than suitable for the needs of game development.

What about the economics? Running AI at scale for millions of players sounds expensive.

Buser: It depends. Let's take a game based in medieval times. We're in a castle, there's several villages to visit, and potentially hundreds of characters. Let's power all those characters with AI and drop a million players into the game. That could be one heck of a cloud bill, especially with the wrong choices. It’s a massive amount of inference at a scale that may overwhelm current business models. Game companies are really struggling with, "I know I can do this stuff, but how do I monetize my game design so I continue to run a profitable organization?"

One cost-control solution is to right-size the models for the experience you're trying to deliver. Building POCs helps you answer these questions. Can I run a much smaller AI algorithm at scale to get the cost down? How does that affect the player experience? What sacrifices will I make? How can I build game designs to mitigate those sacrifices?

We have these conversations with game developers every day. What model is the right model? Where do you run that model? And what types of game designs make sense? Those who answer these questions today will be well prepared for the future. These conversations also underscore how central creator involvement is. For those concerned about the future of work in the AI era, it’s your teams that will be driving the critical decisions that make AI not only effective but affordable.

Looking at the industry landscape, where should CEOs of major game companies focus their attention?

Buser: In the games industry, there are giants, then a middle category, and a very, very long tail of smaller developers. In that long tail are AI-native game companies, founded around novel game design ideas or entirely new development pipelines.

Very large game companies are often caught in the innovator's dilemma. They’re running a giant billion-dollar-plus public company and don’t feel they can afford to take the risk of pivoting a major franchise to an AI-based design. We often work with these companies to weave AI experiences into their existing designs, to begin the process of AI innovation at scale, and prepare for the future.

If you're building the POCs internally, that's great, but you also have to pay attention to the experiences that AI-native startups are building. You have to take it very seriously because while small game companies haven't created the next big hit underpinned by AI, it’s coming. It’s a matter of time, and we cannot ignore it. Put your ear to the ground and start putting AI into production now — even by adding AI experiences to existing games — and you will be innovating a year or two ahead of everyone else.

Opening image created with Imagen 4 and Veo 2 running on Vertex AI, using the prompt: A player playing a console game, illustrated in a flat, colorful style. The characters in the game shift between graphical and game type styles, like shooter, MMO, 8-bit, and other genres, all in one game session.