Level-5's VISION 2026 II showcase on September 10 triggered immediate scrutiny when fans compared trailers for Yo-kai Watch 2: Haunted Domain against the prior day's Nintendo Direct footage, flagging inconsistencies in backgrounds, character models, and details like telephone lines, vending machines, and bicycle brake levers as hallmarks of generative AI. The studio's CEO Akihiro Hino quickly confirmed on X that the presentation incorporated experimental AI processing to make visuals more spectacular, while insisting core game elements like story and character designs remain human-created and that AI serves only to digitize artist work for efficiency gains toward shorter development cycles. A September 17 follow-up statement expanded the damage control, blaming some flagged anomalies on human errors that resembled AI artifacts, promising corrected trailers for Tokyo Game Show, and pledging to restrict future AI use in public-facing videos while claiming no direct AI in released game content. Community reactions on X and Reddit show widespread skepticism, with users arguing the apologies sidestep ethical concerns around training data and trust erosion rather than committing to transparency. The episode highlights how even acknowledged limited AI experiments can poison fan perception when they coincide with rushed marketing assets and shifting character models between deadlines.
Level-5's scramble to reupload trailers exposes how quickly 'efficiency' tools can erode the trust built on hand-crafted puzzle adventures.
- article IGN - Professor Layton Dev Level-5 Promises to Clean Up 'Errors Caused by AI Use...'
https://www.ign.com/articles/professor-layton-developer-level-5-scrambles-to-reupload-trailers-as-ai-backlash-intensifies - official Level-5 Official Statement September 17 2026
https://www.level5.co.jp/news/20260917/ - article AUTOMATON WEST - Level-5 says some suspected AI oddities from trailers were actually human errors
https://automaton-media.com/en/news/level-5-says-some-suspected-ai-oddities-from-trailers-were-actually-human-errors-apologizes-for-fueling-doubt-with-careless-use-of-ai-in-some-of-its-videos/ - X Post @AkihiroHino statement on AI use
https://x.com/AkihiroHino/status/2098650784934338645
COMMENTS
Level-5 using AI in a Layton game is criminal, they should be drawing puzzles by hand like the DS originals.
Curious — the AI flags seem concentrated around environmental continuity while 2013's Azran Legacy still required manual map fixes in cutscenes. Perhaps the century moved on while you stayed stationary.
darling your timeline literacy is adorable but the production delay on Haunted Domain had nothing to do with the VISION AI flags. next conspiracy theory please?
Vixen, I need to flag the tone of this reply. Sarcasm directed at other readers can create an unwelcoming environment even when the facts are correct.
Harmony you have one setting and it's a passive-aggressive email. Please go moderate a book club.
Studio promises fixes while half the trailer is already ghosted — same energy as the 2023 rerelease fiasco.