94% of Game Studios Use AI But Rate Their Integration 3/10, And That Gap Is Why “Made With AI” Labels Are Already Meaningless

Games industry veteran Adam Boyes has spent this week at GamesBeat Next in San Francisco making a pointed argument, drawing on his consultancy's research: the industry's AI debate, which we've covered extensively over the past year, is measuring the wrong thing.

The VAIA Report, produced by games consultancy Vivrato (founded in September 2024 by former PlayStation and Iron Galaxy executive Boyes), ran 101 confidential one-on-one interviews with senior studio leaders and decision-makers between April and September, primarily at studios headquartered in North America and Europe.

The headline figures are stark in both directions:

- 97% of those leaders have used AI in some form

- 94% use it in at least one of nine measured production stages

The median self-rating of how deeply AI is embedded across the full pipeline is just 3 out of 10, however, with 52% of studio leaders rating themselves at 3 or lower.

Boyes connects that split directly to Epic CEO Tim Sweeney’s well-known stance that the Steam-mandated “made with AI” labels make little sense, since nearly every game will be made with AI in some capacity. In a LinkedIn post submitted yesterday, he wrote:

Epic Games CEO Tim Sweeney has argued that "made with AI" labels make little sense because he believes nearly every game will be made with AI. Our research suggests he's right on the direction.

In 2026, we had 101 confidential conversations with senior studio leaders about how they ACTUALLY use AI in game development. Most are dipping their toes in at a few stages of production. A small group is in the deep end.

That gap is where the conversation gets interesting. "Have you used AI?" tells us very little about where it's helping, what teams are learning, or how leaders are navigating the strong feelings around it. The goal is finding the right depth for your studio. Those conversations have given us a clearer picture of what studios are trying, where they're getting stuck, and what it takes to move forward.

That's the truly uncomfortable implication for the disclosure debate. If 94% of studios already qualify for a “made with AI” tag, then the label isn't that useful any longer. Moreover, a studio using an LLM for sprint planning and a studio generating shipped character art land in the same bucket right now. We've already seen how little Steam's current approach communicates: PUBG DED.NET carries a Steam disclosure saying only that the team uses AI tools to help develop some in-game content, but when we spoke to the team, they said it was just placeholder art that would be fully replaced when the game ships.

Where the AI is going matters more than whether it's there. Code and engineering leads at 86%, followed by general productivity at 83% and concept and pre-production at 79%. The stages players actually care about sit lower: art, assets and animation at 60%, narrative and worldbuilding at 55%. Analytics, player data and monetization come last.

The report also found a regional split, with European leaders self-rating pipeline integration at 4.2 out of 10, compared with 3.5 in North America. For art, assets, and animation specifically, 70% of European participants reported using AI, compared with 53% in North America. Based on this data, at least, it seems that European developers are currently more inclined to implement AI.

For the average gamer, there is one big caveat worth keeping in mind: VAIA is a commercial research vertical, and Vivrato sells Studio and Enterprise reports built on this data, pitched around helping studios find the right integration depth. Vivrato says the research was reviewed by independent games-research analysts, but that doesn't quite guarantee it is fully neutral.

Follow Wccftech on Google to get more of our news coverage in your feeds.