Attack on Titan Studio Broke Its Own AI Ban

WIT Studio had a rule. No generative AI in production. The studio behind Attack on Titan, Spy x Family, and Vinland Saga put that policy in place and, apparently, didn't enforce it. On April 4, the first episode of Ascendance of a Bookworm's fourth season aired, and fans immediately started flagging the opening sequence. The backgrounds had that telltale uncanny quality: not quite wrong, but not quite right either. Small distortions. The kind of thing that's hard to unsee once you've seen it.

WIT Studio has since published an official statement confirming what viewers suspected. "We confirmed that [generative AI] was used in the production process of some cuts in the opening sequence," the studio wrote, per machine translation. The kicker is the sentence that follows: WIT states it has "in principle not permitted the use of generative AI in the video production of our works, including this one." So the ban was real. It just didn't hold.

The studio is placing the blame on its own oversight failures rather than on NAM HAI ART, the background art production company involved in the season. That's a reasonable distinction to draw, and credit where it's due for not throwing a subcontractor under the bus publicly. But it doesn't change the outcome: AI-generated frames aired on Crunchyroll, the opening has been pulled from Crunchyroll's YouTube channel, and WIT is replacing the sequence with a fully hand-produced version for all future episodes.

The Problem Runs Deeper Than One Opening

This situation didn't come out of nowhere. Director and animator Terumi Nishii, whose credits include Jujutsu Kaisen and Death Note, posted on Bluesky earlier this year that multiple animation studios are already using AI to generate rough drafts for direction corrections. The specific issue Nishii flagged is that this practice makes it nearly impossible to catch AI usage when scenes are outsourced to other companies. You're not always looking at the raw files. You're looking at finished cuts. By the time something reaches a quality check, the AI involvement can be invisible to anyone not specifically hunting for it.

That's the structural problem here, and it's bigger than WIT. A studio can have a genuine policy against AI and still end up with AI in the final product if the pipeline is complex enough and the checks aren't rigorous enough. As fans spotted on April 4, the distortions were visible to anyone paying attention. The studio's own systems missed it.

WIT reportedly lost 170 million yen in the fiscal year ending May 2025. That context matters. The pressure to deliver content faster and cheaper is real, and it's not unique to WIT. The entire anime industry is running hot, with demand outpacing what traditional production pipelines can handle at the quality level audiences expect. That doesn't excuse what happened here, but it does explain why the temptation exists and why a ban on paper doesn't automatically translate to a ban in practice.

The real question is what WIT does next beyond the apology and the redrawn OP. Pulling the sequence and issuing a statement is the minimum response, not a resolution. If the studio's own production management systems couldn't catch AI usage in an opening sequence, the same gap exists everywhere else in the pipeline. Fans of Attack on Titan and Spy x Family have a reasonable interest in knowing whether those productions are actually covered by the same policy WIT just admitted it failed to enforce.

---

Want to see more? Catch all the latest gaming news, updates, and patch notes right here at XP Gained!