Level-5 Issues Formal Apology at Tokyo Game Show 2026
On September 17, 2026 — the opening day of Tokyo Game Show 2026 — Japanese developer Level-5 published a formal statement on its official website addressing a week-long controversy over alleged AI use in its recent game trailers. The company apologized for "careless use of AI in some videos" and outlined new restrictions on how generative AI will be used going forward.
The timing was deliberate: Level-5 chose TGS's opening day to get ahead of the story while promising to release corrected versions of disputed trailers during the show itself.
Sources: Level-5 official statement (September 17, 2026) · Automaton West coverage · Polygon coverage
How the Controversy Unfolded
| Date | Event |
|---|---|
| Sept. 9, 2026 | Nintendo Direct 9.9.2026 reveals Yo-Kai Watch 2: Haunted Domain trailer |
| Sept. 10, 2026 | LEVEL5 VISION 2026 II: Dreams presentation shows a revised version of the same trailer |
| Sept. 10+ | Fans compare both versions, flagging differences in power lines, vending machines, and the protagonist's face as suspected AI artifacts |
| Sept. 12, 2026 | CEO Akihiro Hino posts on X acknowledging AI use in presentations, insisting the games themselves are human-made |
| Sept. 16, 2026 | Kotaku, Polygon, and other outlets publish detailed coverage of the backlash |
| Sept. 17, 2026 | Level-5 publishes comprehensive apology and explanation on its official website |
| Sept. 17, 2026 | Company announces corrected trailer versions will be uploaded during TGS 2026 |
What Was Actually Disputed
The Yo-Kai Watch 2: Haunted Domain Trailer
The center of the controversy was key artwork and trailer footage for Yo-Kai Watch 2: Haunted Domain, a remake of the 3DS title. Between its Nintendo Direct appearance and its LEVEL5 VISION 2026 II showing, viewers noticed that background elements — specifically power lines, vending machines, and rain gutters — appeared unnatural in ways commonly associated with AI-generated imagery.
Level-5's explanation: those elements were hand-drawn by its designers, but were deliberately simplified because they were background elements. Some differences were also genuine human errors by the artists. The company was not specific about which elements fell into which category.
The reason two different versions existed at all, Level-5 says, is that Nintendo's Direct had a tighter submission deadline, requiring an earlier draft of the trailer to be used. The VISION 2026 II presentation showed the more polished version — hence the visual differences. According to Nintendo Wire, Level-5 confirmed the revised footage will be cleaned up before release.
The Main Character's Changing Face
Between the two trailer versions, the protagonist Nathan Adams (Keita) appeared to have a noticeably different face model. Level-5 addressed this directly:
“"No AI changes or corrections were made to Nathan's face. The character model quality was low, and the director ordered revisions — but because the Nintendo Direct deadline came first, the unfinished model was used there. The VISION 2026 II video shows the updated version." — Level-5 official statement
Professor Layton and The New World of Steam
Some fans also scrutinized trailers for the new Professor Layton title. Level-5 was unequivocal: no AI was used in those trailers, and "there is absolutely no direct use of AI in the game content" of Professor Layton and The New World of Steam.
Level-5's Position vs. Fan Skepticism
What Level-5 Said
The official statement offered three claims and one commitment:
| Claim / Commitment | Detail |
|---|---|
| The disputed background art was human-drawn | Power lines, vending machines, rain gutters — all manual work, per Level-5 |
| Nathan's face change was not AI | Different development stage of the model used in each presentation |
| No AI in game content | Explicit denial for both Yo-Kai Watch 2 and Professor Layton |
| Future AI policy | Limited to internal efficiency only; banned from promotional videos shown to users |
“"We deeply apologize for causing distrust through our immature use of AI and human errors in our productions. Going forward, we will limit AI use to efficiency-enhancing purposes only and avoid direct use in videos delivered to our users." — Level-5 official statement
Where Fans Push Back
The explanation was not universally accepted. As Polygon noted, Level-5 has previously and openly boasted about using AI as a development tool — which means that even when human errors genuinely do look AI-generated, the company no longer gets the benefit of the doubt it once would have. That gap between intent and perception is partly a problem of the studio's own making.
Several commentators pointed to what the apology did not include: CEO Hino did not issue a video apology, the statement did not itemize which specific videos contained actual AI use versus human error, and no independent review of the disputed trailers was offered. For fans who have watched the studio's AI enthusiasm grow in public statements over the past year, the response read as damage control rather than a genuine course correction.
Others took a more measured view: the commitment to ban AI from user-facing promotional videos is a meaningful, enforceable standard — one that other studios have not yet made explicitly.
Why This Matters Beyond Level-5
The Level-5 controversy is part of a broader debate that the games industry has yet to settle: what does responsible AI disclosure actually look like in game development?
Most studios now use some form of AI assistance — in translation, asset generation, QA, or procedural content. The line that fans consistently draw is between back-end efficiency tools and creative work the audience sees. Trailers, key art, and the face of a beloved protagonist are exactly the kind of assets audiences read as representing a studio's craft.
Level-5's misstep — if "careless use of AI" accurately describes what happened — is instructive. When a studio builds a public reputation for AI enthusiasm, even genuine human errors get reinterpreted through that lens. The resulting loss of trust is not easily walked back with a blog post, however accurate the facts in it may be.
What Comes Next
Level-5 has committed to uploading corrected versions of the affected trailers during TGS 2026. The studio's two most visible upcoming titles are now:
- ▶Yo-Kai Watch 2: Haunted Domain — Nintendo Switch 2 (release date TBD)
- ▶Professor Layton and The New World of Steam — Nintendo Switch 2, Nintendo Switch, PlayStation 5, Steam (release date TBD)
Neither title has a confirmed release window, and neither has been officially announced for regions outside Japan at this stage.
GamePeak Take
Level-5's statement resolves the factual dispute — power lines were hand-drawn, Nathan's face changed for normal production reasons — while confirming that AI did find its way into some of the company's public-facing materials without adequate transparency. The new policy banning AI from promotional videos is a concrete step, and it sets a bar that fans can hold the studio to.
What it does not resolve is the trust deficit built up by months of the studio publicly framing AI as a tool for making games faster and cheaper. That deficit will be paid down through actions, not statements. TGS 2026 is where the corrected trailers land; what the games themselves look and feel like at launch will tell a clearer story than any blog post.
The harder industry question — how studios should proactively disclose AI use before fans start auditing trailers frame by frame — remains unanswered. Level-5 stumbled into providing a case study for that conversation whether it wanted to or not.