The GDC 2026 State of the Industry survey asked 2,300 game professionals what they think about generative AI. The results are contradictory in a way that tells you more about the state of AI tooling than any product launch or benchmark ever could.
52% said AI is bad for the game industry. That number was 30% last year and 18% the year before. Only 7% called it positive, down from 13%.
In the same survey, 36% said they use AI tools at work. Among those users, 81% use AI for research and brainstorming, and 47% use it for code assistance.
Developers are increasingly convinced AI is harmful. They're also increasingly using it. What's going on?
The sentiment split isn't random
The GDC 2026 report breaks down who hates AI and who uses it, and the pattern is clear.
Negative sentiment by discipline:
| Role | "AI is bad for the industry" |
|---|---|
| Visual / technical art | 64% |
| Game design / narrative | 63% |
| Programming | 59% |
AI adoption by seniority:
| Level | Uses AI at work |
|---|---|
| Upper management | 47% |
| Lower-level employees | 29% |
| Publishing / support firms | 58% |
| Game studios | 30% |
The people who think AI is bad are mostly the people whose work AI threatens to replace: artists and writers. The people using AI the most are managers and publishing staff who use it for emails, reports, and brainstorming. These are largely different groups having different experiences.
One surveyed developer put it plainly: "Why would I replace human creativity with a regurgitated amalgamation of everything that's come before?"
What developers actually use AI for
The use case breakdown from the same survey is revealing:
| Use case | % of AI users |
|---|---|
| Research / brainstorming | 81% |
| Daily tasks (emails, code) | 47% |
| Asset generation | 19% |
| Procedural generation | 10% |
| Player-facing features | 5% |
The overwhelming majority use AI for research and code. Almost nobody ships AI-generated content to players. The industry has collectively decided that AI is a backstage tool, not a front-of-house feature.
This matches what most developers I've seen in forums and postmortems describe: AI is useful for scaffolding code, debugging, and getting past blank-page paralysis. It falls apart when you need pixel-perfect art consistency or narrative voice.
The tool preferences are lopsided
ChatGPT dominates at 74% adoption among AI-using developers. Google Gemini sits at 37%, Microsoft Copilot at 22%. Text-to-image tools lag far behind: Midjourney at 17%, Adobe Generative Fill at 13%.
The gap between LLM adoption (74%) and image generation adoption (17%) tells you where AI actually delivers value in game dev. Writing and debugging code with a language model works well enough to become habitual. Generating game art that fits a specific style, pipeline, and resolution requirement is much harder.
This is also why tools that understand game engine context specifically (like Ziva for Godot, or the various MCP-based tools connecting Claude to game editors) have found a niche. A general-purpose LLM generates generic code. A tool that reads your scene tree and project structure generates code that fits.
28% lost their jobs in two years
The AI sentiment doesn't exist in a vacuum. The GDC survey found that 28% of respondents lost their job in the past two years. 74% of students said they're concerned about their future job prospects.
When over a quarter of your colleagues have been laid off and AI is being positioned as the reason companies can "do more with less," negative sentiment isn't irrational. It's self-preservation. The developers who view AI most negatively (artists at 64%, designers at 63%) are the ones whose roles are most visibly threatened by generative models.
Where this leaves us
The data paints a picture of an industry with two parallel realities. In one, developers quietly use AI to brainstorm, scaffold code, and automate repetitive tasks. In the other, the same developers watch their colleagues get laid off while executives pitch AI as a cost-cutting measure.
Both realities are true simultaneously. AI tools do save time on specific tasks. AI rhetoric is also being used to justify smaller teams and lower budgets. The 52% who view AI negatively aren't technophobes. They're reacting to how AI is being deployed as a labor strategy, not just a productivity tool.
For developers building AI-assisted tooling (and I work on one), the takeaway is straightforward: the tools that survive will be the ones that make individual developers faster without making their employers think they're replaceable. That means augmenting decisions, not automating them. It means context-aware assistance within the editor, not "AI generates your game for you" marketing.
The 36% who use AI tools despite the backlash aren't hypocrites. They've found narrow, specific tasks where AI actually helps. The 52% who view AI negatively aren't wrong either. They're responding to a real threat that has nothing to do with the technology and everything to do with how companies choose to use it.
Top comments (0)