Five identical references. One very predictable result.

Original work needs a brief and an approval process that can tolerate an unfamiliar answer.
Ask for something original, supply five near-identical examples, then reject the unfamiliar proposal because there is no successful reference for it. The brief has just argued itself into a corner.
AI can make the consequences arrive faster. It also gives a team a convenient suspect when the work looks like everything else. The choices around the tool deserve the same scrutiny as the output.

Similarity can begin before production
Research on AI-assisted storytelling points to a genuine risk. Wan and Kalman's study describes earlier findings in which AI ideas improved individual stories while reducing diversity across the group. Their own experiment used varied AI personas to generate inputs and found that story diversity could be preserved against a human-only baseline. These are findings about storytelling, with limits on how far they can be applied to design. Wan and Kalman's research.
For a design team, that suggests a question worth testing: how much variety is being allowed into the process before anything gets selected?
If everyone sees the same polished reference first, it can become the standard every later idea has to satisfy. That is an approval problem as well as a production choice.

Start with something the business actually owns
Ask each person to bring an observation before showing the generated directions. It could come from the product, a recurring customer question or a detail in how the service works. Keep private customer information out of tools that have not been approved to handle it.
Use those observations to judge the ideas. “It looks premium” leaves too much unexplained. What makes this direction belong to this business? Which part of the offer does it help a customer understand?
Try replacing the brand name with a competitor's. If the idea still works without a single adjustment, investigate what is missing.

Leave room for a rough proposal
A familiar direction arrives with evidence that someone else made it attractive. An unfamiliar one may still need a sketch or a material test. Give it a fair chance to become clear before comparing its finish with a completed campaign.
References can guide a discussion. They should not become a requirement that every new idea arrive with somebody else's approval stamp.








