Artificial intelligence is making significant advances in the gaming industry by streamlining development and creating more immersive experiences, from reactive characters to dynamic environments and personalized gameplay.

At Game Developer Conference 2025 today, Google Cloud revealed how AI startups and AI-native gaming studios are using its AI architecture to build games in new ways to overcome the biggest challenges in the industry.

To understand how the gaming industry is evolving, SiliconANGLE spoke with Jack Buser, director for games at Google Cloud, and executives at AI gaming startups about how the industry is changing and what developers and players can expect.

“First and foremost, there’s AI for game development, and then there’s another category, which is AI for new player experiences,” Buser said. “We saw game companies start to put generative AI into actual production pipelines over a year ago now.”

In many cases, game studios can use generative AI to handle mundane tasks that consume development time, freeing up developers and creative workers to focus on more innovative aspects of game design.  “We see game companies actually use this to accelerate their game development cycles,” he explained.

Building 3D models and testing games with generative AI

One company helping gaming studios accelerate their development cycles is Common Sense Machines Inc., which builds state-of-the-art generative AI models and agents that let users produce controllable, production-ready 3D artwork from images, text and sketches.

“Creating anything 3D can take hours or days, even for professionals right now,” Tejas Kulkarni, co-founder and chief executive of Common Sense Machines, told SiliconANGLE. “A single asset, like a chair, could 16 hours to make before you can see it on a website or in a game or in an industrial simulation platform.”

The company describes its tool as a “3D copilot” that lets users upload an image or a sketch, get their 3D model, fine-tune it and then export it into their development workflow.

Common Sense Machines works with an AI-native gaming studio named Cosmic Lounge, powered by Google Cloud’s fully managed model development and engineering platform Vertex AI. CSM’s service helps generate visually engaging promotional content for marketing, develop seasonal decorations for in-game content, such as Christmas-themed pet designs, and rapidly refresh game content with new variations to keep players engaged.

Testing can be tedious and slow, requiring long hours repeating the same boring tasks. That’s why Nunu.ai created multimodal AI agents with Google Gemini models that can navigate through games similar to human players to identify bugs to reduce tedium and burnout for testers.

“Why AI engines? It mostly comes down to that games are interactive environments. So, it’s very hard to write a test script because it changes,” Nicolas Muntwyler, co-founder of Nunu.ai told SiliconANGLE.  “Writing a script sometimes doesn’t work. If there’s suddenly a new event, a new pop-up and it just breaks, that’s a lot of maintenance.”

Using Nunu.ai’s AI agents, a tester can simply give the agent a goal using natural language and let it explore the game. It will then play like a real person — using items, shooting guns, crafting and interacting with the user interface — just as a player would. These are the kinds of tasks a tester would need to perform whenever a game version or function changes. However, they are also the most monotonous parts of game testing.

A tester could ask the AI to complete a specific task, such as crafting an axe or playing through a chapter of the game. Muntwyler noted that the AI is slightly slower than a human or script because it mimics how a person would naturally play. If it succeeds, it provides a pass checkmark; if it encounters an error, it reports the issue, describes how to reproduce the bug, and provides a visual trace so testers can follow along.

“What we want to do is take the large tester ‘to-do list,’ automate as much as we can, and free up testers to do testing that is valuable,” said Muntwyler. “In my opinion, that’s testing if the game feels fun to play or not, something AI can’t do. That’s exactly where humans would shine and they could do a great job at it.”

The emergence of ‘Living Games’

Series Entertainment Inc., an AI-native game studio, has been putting Google Cloud’s AI infrastructure to streamline its content production and enable its developers to build games faster than traditional teams. It’s also using AI to drive reactive gameplay experiences that adjust to what players do.

“We coined a term which drives our strategy, called ‘Living Games,’” Buser said. “This represents what happens when these worlds collide. You’re actually able to not just transform what a game company and game development studio can look like and how they behave, but also you can transform the player experience.”

Series Entertainment uses AI to free up its development team from repetitive tasks that take away time from more creative tasks to allow them to find time to tell stories. The studio also intends to build on the “living games” aspect by building generative AI directly into the gameplay loop by adding it to the game mechanics.

“Something we’re going to be releasing next year [is] we have AI built into the game mechanics that interact with the player to help tell the story,” Josh English, chief technology officer of Series Entertainment, told SiliconANGLE. “So based on the player’s interactions with the game environment itself, it can adapt to tell an immersive story together.”

He said if you look at an ordinary game studio, a great deal of work is put into producing new content for a game to keep the player base happy. Adding AI to the equation, it can tailor the game and gameplay experience on the fly.

“It radically reduces the cost involved and the time involved in producing new content,” English added. “At the same time, on the back end, it’s still empowering our creators to build new experiences as well. So, I think it’s bringing the best of both worlds.”

Image: Series Entertainment