Games Workshop:
A clear policy on banning AI.

Games Workshop’s CEO has banned generative AI from the company’s design process. Not an IT decision. Not a compliance decision. A commercial decision.

Why? Because their business model depends on protecting distinctive intellectual property. And AI-assisted content may be harder to protect as IP. Competitors may be able to legally copy it.

That’s governance working as it should: identifying what drives value, and managing the risks to it.

Many AI policies start from the wrong place. They start from the tool (“is ChatGPT allowed?”) or they become lists of standard risks (data security, privacy) and end up as compliance exercises.

The better approach starts from your strategic objectives.

• What are you trying to achieve?
• Where could AI help or hinder that?
• What boundaries protect value while still enabling progress?

For Games Workshop, it’s protect their IP-based business model. So no gen AI in design. Clear boundary.

For other organisations it might be different. If speed to market is critical, AI may be essential in parts of the process. If customer trust depends on human judgement, AI may need boundaries there.

The policy follows from the objectives and risks. Not the other way around. (Worth reading the Pinsent Masons article on this, which sets out the IP uncertainty clearly.)

At Brave, we call this Objectives@Risk: a clear line from what you’re trying to achieve, through the risks to those objectives, to the controls that actually matter.

Games Workshop’s stance is set out in a couple of paragraphs in their half-year report. That feels about right, when you’re clear on what matters, the rules can be short.