How you're rewarded shapes whether AI makes everyone think the same

When people are paid for originality instead of quality, they use AI differently — they cherry-pick suggestions instead of accepting them wholesale. This means the homogenizing effect of AI isn't inevitable. It depends on the incentives around how you use it.

The worry about generative AI is that it narrows human creativity because everyone leans on the same tool the same way. This experiment shows that's not baked into the technology. The actual narrowing happens upstream, in how people are told to succeed. A company, school, or team that rewards only polish and speed will get uniform outputs. One that rewards actual novelty gets diversity — the AI just becomes a different kind of tool. The implication is uncomfortable: if your organization is seeing AI homogenize everything, that's partly because you built it to.