Credit: Unsplash/Glenn Carstens-Peters. The video game industry is in the middle of a revolution. Artificial intelligence is no longer just a behind-the-scenes helper for tweaking code or creating concept art. It’s now generating voices, environments, characters — and even replacing human testers.

That shift is unfolding fast. A report from Totally Human Media shows that the number of games on Steam disclosing AI use rose 700% in just one year — from about 1,000 titles to nearly 8,000. That’s about 7 percent of the entire Steam library, with almost 20 percent of new releases this year openly admitting to generative AI in their development.

Studios see AI as a way to cut costs and speed up production. But the shift is stirring excitement and dread in equal measure.

From Concept Art to Entire Ecosystems

When ChatGPT first rolled out in late 2022, many studios used it for relatively small tasks: generating simple dialogue or brainstorming concept art. It felt like an experimental sidekick, not a replacement. But by spring 2025, the scope had exploded.

At the Game Developers Conference in San Francisco, engineers from Google DeepMind showed off autonomous agents that could play through early builds of games, spotting glitches without the need for human testers. “Autonomous agents” could do in hours what teams of QA workers once spent weeks grinding through.

Microsoft unveiled a system that could watch a short video and generate entire level designs and animations from it — processes that normally take hundreds of painstaking hours. On the same stage, Roblox executives presented Cube 3D, a generative AI tool that can spin up fully functional objects and environments in seconds from a simple text prompt.

Roblox asset made with Cube 3D. Prompt: “A red buggy with knobby tires.” These demos signal a future where studios might lean on AI for the most expensive, time-consuming parts of production. For an industry still reeling from repeated layoffs and spiraling costs, the appeal is obvious.

The Surge in AI-built Games

The Totally Human Media report adds that among studios that admitted using at least some AI in their titles posted on Steam, around 60 percent of games used AI for visual asset generation — characters, textures, and backgrounds. Others turned to large language models for voices, background music, or branching story arcs. Some developers even relied on AI for marketing copy, offensive content moderation, and coding support.

Popular titles already on the market prove the trend isn’t confined to indie experiments. My Summer Car features AI-generated paintings, Liar’s Bar uses AI for character voices, and The Quinfall employs AI-made interface images. The first of those has sold over 2.5 million copies.

AI-generated paintings featured in the video game My Summer Car. Credit: Reddit.

NPCs Who Talk Back

The dream of intelligent non-player characters has haunted gaming for decades. Now, AI is delivering the goods. Nvidia recently teamed up with startup Convai to create NPCs in a cyberpunk ramen shop who can carry on improvised conversations with players. *PC Gamer*shared clips of the experiment, and the interactions feel closer to roleplay with another person than the canned responses of traditional games.

Sony has been testing something similar. By combining OpenAI’s speech recognition system with its own tools, the company created an AI-powered version of Aloy, the heroine of Horizon Forbidden West. Instead of delivering pre-recorded lines, Aloy can answer player questions in real time.

And then there’s the boundary-pushing work from researchers at Google and Stanford. In late 2023, they developed “generative agents” as proxies for human behavior—virtual characters that mimic human routines and behaviors. According to their report, “Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day.”

In their simulated town inspired by The Sims, these characters didn’t just walk through routines. They noticed each other, formed relationships, and even organized a Valentine’s Day café date without human input. The experiment looked less like a video game and more like a simulated society of digital beings.

Not Just Talk

Square Enix, the company behind Final Fantasy, invested in Vienna-based startup Atlas to test AI-generated assets. Atlas also partners with the studio Parallel, which recently previewed a game where players can create AI-generated armor for their characters.

Other startups aim higher. Decart, led by founder Dean Lietersdorf, is training “world models” that generate explorable environments from a perspective similar to first-person shooters. Google and World Labs — founded by Stanford AI pioneer Fei-Fei Li — are also developing similar systems.

For Lietersdorf, cost is the main hurdle. Google’s Veo 3 model costs about $1,000 to generate an hour of video. Decart’s smaller models can cut that to 25 cents an hour, though at lower quality. Lietersdorf believes the magic number is 10 cents an hour.

“Once AI models can reliably generate high-quality video at less than 10 cents an hour, game studios can offer such games on a subscription basis.”

“The entire gaming industry will change when there‘s a single game that’s completely AI developed, and that game goes very viral, and I think we’ll probably see that in the next three to six months,” he added during an interview with The Information.

A Revolution, But Not Without Resistance

Without video games, the artificial intelligence boom might never have happened. The Nvidia chips that powered recent AI breakthroughs were first designed to render video game graphics. Now, game studios are returning the favor by using generative AI to automate some of the hardest, most expensive work in game development.

Jacob Navok, CEO of Genvid, is blunt about the inevitability of the gaming industry’s shift to AI: “In the same way that digital filmmaking replaced people painting on cels, this is going to happen,” he told The Information. His company uses AI for cut-scene animation but they had to make their own tools as public models struggle with visual consistency frame by frame. For now, humans still outperform machines in areas like music, editing, and voice acting, Navok says.

Navok recalls that producing 20 minutes of animation once cost around $2 million. With new AI video-generation tools from Google, MiniMax, and Kling, the same job now costs about $1,500. Time savings are just as staggering. Navok said one deputy took just six hours to build three minutes of a prototype game — a task that previously required six months.

The stakes are high. Studios are still reeling from mass layoffs and the financial burden of making hyperrealistic games (that’s what most gamers want nowadays) that sometimes lose money despite huge sales. AI is being pitched as the answer — cheaper, faster, and endlessly adaptable. But some analysts warn that pouring resources into generative AI might distract from deeper structural problems in the industry, such as high inefficiency.

What once took millions of dollars and months of labor can now be done in hours for a fraction of the cost. The tools are still rough, but they’re improving fast.

For gamers, the near future could mean unpredictable worlds full of characters who feel alive in ways we’ve never seen before. For developers, it’s a gamble between survival and surrendering creative control to machines.