We’ve seen a slew of companies clarify their stance on generative AI over the past few months, and now it’s Capcom‘s turn to draw a line in the sand.
“Our company will not implement the materials generated by our AI into game content,” Capcom stated in a summary published from its February 2026 investor briefing session. The publisher added that it does plan to “proactively” use AI tools as a “contributing technology to improve the efficiency and productivity of the game development process” — specifically calling out graphics, sound, and programming as areas currently under exploration.
How all of this folds into the actual development pipeline remains unclear, but it’s worth noting that back in Summer 2025, Capcom’s technical director Kazuki Abe had already been experimenting with generative AI in collaboration with Google — specifically using Gemini Pro, Gemini Flash, and Imagen to help streamline the “idea generation” process for in-game environments. A summary of that investor session is available here for anyone curious. The key idea at the time was that coming up with the “thousands to tens of thousands” of unique concepts needed for even minor in-game objects is one of the most labour-intensive parts of development — and that gen-AI could help shoulder that load before human artists take over.
To put it simply: Capcom will use generative AI to help with time-consuming, repetitive groundwork, while keeping finished assets in human hands. The catch? Other developers are already trialling similar approaches, and AI-generated assets have a habit of slipping through QA and ending up in front of players anyway. Crimson Desert developer Pearl Abyss, for example, recently confirmed that some assets created using experimental AI generative tools were never swapped out of the game’s launch build — the latest in a line of studios including Ubisoft and 11 Bit Studios that have issued similar apologies.
So while Capcom is nominally drawing a clear boundary, its position is slightly undercut by timing. Just last week, Nvidia’s DLSS 5 reveal used Resident Evil Requiem as its showcase — and the backlash was swift. DLSS 5’s neural rendering noticeably altered the appearance of protagonist Grace Ashcroft, softening her features into what players immediately dubbed a “yassification filter.” What made things worse is that reports suggest Capcom developers found out about the DLSS 5 demo at the same time as the public, raising concerns internally about the studio’s anti-AI reputation being damaged by association.
Nvidia CEO Jensen Huang pushed back, insisting that critics were “completely wrong” and that DLSS 5 output remains “in the direct control of the game developer.” Capcom executive Jun Takeuchi had publicly endorsed the technology in Nvidia’s press release. The contradiction — signing off on Nvidia’s AI-powered alterations to Resident Evil Requiem‘s artistic vision, then days later clarifying that AI-generated assets won’t feature in Capcom games — sends a message that depends a lot on where exactly you draw the line between an “asset” and a post-processing layer.
Meanwhile, publishers like Square Enix are taking a far more aggressive approach. Square Enix has openly committed to automating 70% of its QA and debugging tasks using generative AI by the end of 2027 — a goal that has drawn its own share of controversy around potential job losses. Capcom, by contrast, is trying to carve out a more moderate position: AI in the pipeline, but not on screen.
Whether that distinction holds up in practice — especially given how easily placeholder assets have escaped into shipped games elsewhere — is the question the industry will be watching closely. Don’t expect the generative AI debate to go away any time soon.
Speaking of 2026 releases navigating their own turbulence, if you missed what’s been going on with the Call of Duty: Modern Warfare 2019 Steam spike that’s got the gaming world talking, it’s worth a read.
