Studios formalize AI policies as grievance workload and data risks mount
Summary
A straw poll at GamesIndustry.biz's HR Summit found most companies now have an AI policy.
That marks a shift from a year earlier when only a minority had one.
HR teams report rising grievance letter workload tied to employee generative AI use.
Prompt injection risks forced policy rewrites within weeks of launch, undercutting apparent progress.
01
Policy adoption accelerates while control lags behind
This year's straw poll found most companies now have a formal AI policy in place.
One leader rewrote AI guidelines two weeks after learning about a new prompt injection risk.
Enterprise AI licenses limit vendor-side data training risk but miss personal-account use entirely.
Policies now function as moving targets given this pace of change.
02
Grievance letters inflate HR workload and legal risk
A legal representative said they receive about ten AI-written grievance letters weekly from clients.
One nine-page grievance contained eight pages of claims the HR team could not interpret.
Tribunals can dismiss cases when employees cannot explain their own AI-generated submissions.
HR headcount has shrunk while this letter volume keeps workload flat or rising.
"Quite often it's given me more to do."
03
Unsupervised AI pilots erase their own time savings
An artist used Gemini-generated shader code while the usual programmer was away.
The code broke coding standards and went unnoticed for several weeks before reversal.
A freelancer's AI-generated art asset shipped with extra fingers before the community reacted.
Most AI pilots collapse because workflows rely on individual tacit knowledge instead of documented procedure.
04
Personal AI accounts bypass company data governance
Employees can paste NDA-covered material into personal ChatGPT accounts outside company oversight.
One leader said there is no way to physically stop this individual behavior.
Several leaders favor clearer acceptable-use guidance over technical enforcement as the practical fix.
05
Most studios remain early in AI adoption maturity
Most companies still sit at the first of four defined AI adoption stages.
That stage typically means buying an enterprise license for coders to experiment with AI.
Only larger studios have reached stage two, piloting AI inside defined production workflows.
AI speeds coding and automates mundane tasks without producing original creative results, one senior leader said.
"I haven't seen AI create anything fun."
What this means
For Publishers & Developers: Add AI-disclosure clauses and human sign-off gates to every freelancer and outsourcing contract.
For Service Vendors: Pitch guaranteed AI-disclosed, human-QA'd art delivery to publishers tightening external asset contracts.
For Investors & VC: Add AI governance maturity and NDA-leak exposure to studio diligence checklists.
Events
Companies
—
Games
—
Locations
—
COMPILED BY GAMES ATLAS EDITORIAL
RELATED ARTICLES
EXPLORE MORE
Comments (0)
No comments yet. Be the first to share your thoughts!