Two years ago, generative AI in game development was mostly a talking point. In 2026, it’s a production line item — and the industry is deeply divided about whether that’s a good thing.

The adoption numbers

According to the GDC 2026 State of the Industry survey, 52% of studios now use generative AI in production — not experimenting on the side, but actually shipping games with AI somewhere in the pipeline. That figure has roughly doubled in two years. On Steam alone, 7,300 games disclosed the use of generative AI assets in 2026.

The reason isn’t mysterious: asset-heavy games get more expensive to make every generation, and budgets haven’t kept pace. A background prop that used to take an artist four to eight hours can now be generated in minutes.

Where AI actually gets used

The survey breaks down use cases, and the pattern is telling — AI is doing the most work in the least visible parts of development:

• Research and brainstorming — 81% of studios
• Coding assistance — 47% of studios
• Prototyping — 35% of studios
• Drafting initial design documents — 28% of studios, using AI to generate first-pass structure, quest frameworks, and narrative threads for human designers to refine

Individually, 36% of game industry professionals say they personally use generative AI tools at work, with ChatGPT the clear leader (74% adoption among AI users), followed by Google Gemini (37%) and Microsoft Copilot (22%). Major players — Google, Roblox, Unity, Ubisoft, and Microsoft among them — are now investing directly in AI-powered game creation tooling rather than treating it as a third-party add-on.

The part everyone skips: over half the industry doesn’t like this

Here’s the number that gets buried under the adoption stats: 52% of industry professionals believe generative AI is having a negative impact on the industry — essentially the same share of studios that are using it. Adoption and skepticism are running at almost identical rates, which tells you AI use in games right now is less “everyone loves this” and more “everyone feels they can’t afford not to.”

That tension is likely to define the next few years more than the technology itself. Faster, cheaper asset production is real. So is a workforce watching layoffs climb in the same period AI tooling adoption doubled — a connection that’s hard to prove directly, but not hard to understand why developers are drawing it.

---

Speaking of layoffs — see the actual scale of 2026’s job cuts against the industry’s record revenue. Or check out how indie developers are using the same tools very differently than the AAA studios above.