Level-5's Akihiro Hino Backs GenAI To "Enrich the Creative World" - What That Means for Creatives
Level-5 president Akihiro Hino is going public with a strong take: generative AI can enrich creative work, not dilute it. Fresh off Inazuma Eleven: Victory Road and Fantasy Life i: The Girl Who Stole Time, he says many major studios are already using GenAI-disclosed or not.
His bold claim: the right AI workflows could compress AAA timelines from five-ten years down to about two. His goal is simple-games that feel like dreams realized, built faster without sacrificing ambition.
Why this matters for your craft
Shorter cycles mean more iteration, more experiments, and less time sunk into grunt work. The leverage shifts to taste, direction, and IP-areas where human judgment is the moat.
- Concept exploration: rapid style frames, mood boards, and variations to spark art direction.
- Blockouts and prototypes: faster greybox assets and placeholder VFX to test ideas early.
- Narrative and VO: draft passes for barks, item flavor text, and placeholder VO for timing.
- Localization assists: rough passes that native linguists refine and approve.
- QA triage: clustering duplicate bugs, log summarization, and content compliance checks.
- Tools and automation: quick scripts, shaders, or editor utilities to remove repetitive tasks.
The pushback-and the fix
The biggest concern: models trained on creators who never consented. There's also frustration around infrastructure strain, including the recent RAM crunch many blame on runaway demand from major AI players.
Hino's stance is that branding AI as "bad" stalls progress. The healthier move is to set guardrails that respect creators while keeping teams competitive.
- Use licensed, consented, or in-house datasets; avoid scraping unlicensed work.
- Track provenance with asset logs and content credentials (see C2PA).
- Fine-tune on studio-owned assets; get explicit consent for any client or contractor work.
- Install a human review gate before anything ships; credit AI-assisted roles where used.
- Offer opt-in/opt-out options for contributors and vendors.
A lean workflow you can ship with
- Preproduction: define constraints, style guides, and references; generate wide, choose narrow; hand-finish.
- Production: use AI for placeholders and drafts; tag every AI-touched asset; replace with final art as it's approved.
- QA and polish: AI for surfacing patterns and regressions; humans for judgment calls.
- Legal and ethics: scheduled dataset reviews, model cards, and creator consent audits.
Where this lands
GenAI is a tool, not a vision. Teams that pair strong taste with clear standards will win on quality and speed. The rest will look and feel generic.
Hino puts it plainly: "I want to see and create games that go beyond today's AAA standards… I hope both creators and audiences can recognise AI as a tool that people use to create their work."
If you want a deeper look at the studio behind the stance, explore Level-5. If you're mapping AI skills to your role, here's a curated path: AI courses by job.
Your move
Could GenAI be used ethically in game development-or will it stay a hinderance? Where would you draw the line in your pipeline?
Your membership also unlocks:
