Dead Space and Callisto Protocol veteran Glen Schofield says he’s “100% behind AI” for game development. In a new interview, he likened today’s AI moment to the arrival of Photoshop and motion capture – tools that initially spooked creatives but ultimately raised the bar. That stance lands amid a real adoption curve: the latest GDC survey notes 52% of studios already use gen-AI tools, even as curiosity cools and some companies push back.
Where Gen-AI Helps Today (Not Hype)
Start with sanded, scoped use cases that measurably reduce toil without eroding IP quality or team morale. By identifying specific areas where efficiency can be improved, you create a foundation for a more productive workflow. This approach ensures that the entire team remains focused and motivated, as they can see tangible results from their efforts.
| Workflow Area | Concrete Use Case | Guardrails | KPI to Track | Typical Win |
|---|---|---|---|---|
| Pre-production art | Mood boards, variation studies, silhouettes | No “finals” from AI; human paint-over required | Concepts/day, director revision cycles | Faster alignment on look |
| Narrative tools | Beat-sheet exploration, alt barks, quest stub passes | Writers review + rewrite; style guide prompts | Time to first draft, % accepted lines | More options with same headcount |
| Audio prototyping | Temp SFX variants, VO scratch | No synthetic VO in ship build; actor replacement policy | Iteration speed, handoff time | Quicker feel checks |
| Code assist | Boilerplate generation, unit-test scaffolding | No secrets in prompts; mandatory code reviews | PR cycle time, bug density | Faster scaffolds, fewer regressions |
| QA & UX | Test-case generation, log summarisation | Human triage; no auto-closing | Bugs found/test hour, repro rate | Wider edge-case coverage |
Insider Tip: Move one pipeline at a time from manual to AI-assisted to human-finished. Publish before/after metrics internally to build trust.
Risk Map: What Can Go Wrong (and How to Prevent It)
Adoption without policy is how studios get burned. Use a lightweight risk mitigation plan so you can scale without headlines or headaches.
| Risk | What It Looks Like | Mitigation Playbook | Owner |
|---|---|---|---|
| IP contamination | Model trained on unlicensed art | Vendor attestations; “no training” clauses; internal model for sensitive work | Legal + Procurement |
| Style drift | Assets feel generic or off-brand | Locked style guides; reference bundles; art director approvals | Art Direction |
| Privacy & leaks | Build logs or secrets in prompts | Redacted prompt layer; on-prem or VPC tools | Security + Tools |
| Credit & morale | Creators feel replaced | Credit policy (“AI-assisted by…”), keep humans on final pass | Production |
| Over-automation | Shipping placeholder quality | Ship-blocker QA gates; minimum “hand finish” % | QA + Leads |
Insider Tip: Publish a 1-pager AI Use Policy. If everyone can find it and it’s plain language, you just saved twenty meetings.
Playbooks by Studio Size
Different teams need different AI cadences, as each function and project brings its unique set of challenges and requirements. Understanding these variations is essential for maximising the effectiveness of AI integration within an organisation.
| Studio Type | 90 – Day Plan | 6 – 12 Month Plan | What Not to Do |
|---|---|---|---|
| Solo / Tiny indie | Adopt code assist + image variation; one clear policy | Fine-tune tiny local models for style refs | Don’t outsource final art or VO |
| AA (20–80) | Pilot AI in concept, test plans, scripting | Central prompt/style library; add on-prem tooling | Don’t touch core player-facing voice/quest finalisation |
| AAA (80+) | Create an “AI Ops” tiger team; vendor review | Internal eval bench; data governance; training | Don’t allow shadow tools – centralise |
| Live-service | AI for patch notes, telemetry clustering, CS macros | Automated test-case expansion; triage summarisation | Don’t let AI “talk as the brand” |
| Mobile F2P | Campaign copy variants, ad iteration, store A/B | LTV uplift experiments with guardrails | Don’t over-personalise without consent |
Insider Tip: Name an AI Product Owner (not just an enthusiast). If it’s part of someone’s job, it actually happens.
Skills Matrix: Level Up Your Team (Don’t Shrink It)
AI shifts what we do, not why. By leveraging its capabilities, we can enhance our creative potential and push beyond our current limitations. Use it to raise the creative ceiling while trimming grind, allowing us to focus more on inspiration and innovation rather than repetitive tasks.
| Role | AI-Aided Tasks | Human-Only Value |
|---|---|---|
| Concept Artist | Rapid comps, palette ideas | Taste, cohesion, IP authorship |
| Writer/Narrative | Alt lines, quest scaffolds | Voice, theme, subtext, continuity |
| Designer | Enemy/balancing stubs | Feel, pacing, experiential tuning |
| Engineer | Boilerplate/tests | Architecture, performance, review |
| Producer | Risk/roadmap drafts | People, negotiation, prioritisation |
| Audio | Temp SFX sketches | Emotion, mix, performance direction |
Insider Tip: Rebrand adoption internally as “remove toil, keep taste.” It’s the right ethos and sells the change.
Final Thoughts
Schofield’s “100% behind AI” isn’t a call to replace creatives; it’s a push to weaponise tools against spiralling scope and cost. The winners won’t be the studios that automate the most – they’ll be the ones that measure, govern, and direct AI to unlock more human craft in the same calendar. Adopt it like infrastructure. Pilot where it’s boring. Guard what’s sacred. And make your game feel more authored, not less.
That’s it for this one! Please like, share, and comment if enjoyed this article AND…
Grab Myles’ FREE ebook now and find 15 indispensable design patterns that will equip you to craft exceptional Web3 gaming experiences. We’ll also notify you when his new book on immersive design is out!