Our AI policy is a shrug wearing a lanyard

A fictional team meeting ends with permission to “use AI responsibly.” Everyone nods. The phrase has the useful specificity of “have a normal Tuesday.”

People need to know which tools are approved and what information can be entered into them. They also need a named person to ask when a client file or unfamiliar use case creates uncertainty. Enthusiasm is not an information-handling policy.

Write down a few actual examples from the team's work. Is generating a rough internal concept acceptable? What review is needed before an output reaches a client? Who checks details, permissions and the final production file?

Give those checks time in the schedule. If a new tool makes first drafts faster, that does not mean verification has evaporated. Someone still signs off on what leaves the studio.

Review the arrangement as the tools and client requirements change. A document nobody can find is merely a policy with excellent camouflage.

The lanyard may remain. At least it can hold a name badge for the person who is responsible.