AI makes each of your people more creative. It quietly makes all of them more alike.

Hand your team generative AI and the work gets noticeably better. Drafts are cleaner, ideas arrive faster, and even your less confident people start producing sharp, polished, creative-looking output. Then, a few months in, you notice something harder to place: the decks all sound the same, the campaigns rhyme, the "fresh" ideas are fresh in the same way. Everyone got better, and somehow the collective output got more uniform. That is not your imagination, and it is not a culture problem. It is a measured effect of the tool.

A 2024 study in Science Advances put it to the test. Writers given access to generative-AI story ideas produced work that was rated more creative, better written and more novel than those working alone, and, tellingly, the largest gains went to the least naturally creative writers. AI genuinely lifted individual creativity. But the same study found that the AI-assisted stories were more similar to one another than the stories humans wrote unaided. The authors name the trap precisely: a social dilemma in which, with generative AI, writers are individually better off while collectively a narrower scope of novel content is produced. The tool raises the floor and, in the same motion, pulls everyone toward a common middle.

Why would "more creative" also make everyone more alike?

Because everyone is drawing their ideas from the same models, trained on the same data, optimised to produce the strong, plausible, on-distribution answer. That is exactly what makes AI so useful to an individual: it hands you a good default. And it is exactly what homogenises the group, because a good default is, by definition, close to the average of everything that came before. The very quality that helps the single writer, a confident, sensible suggestion, is what converges the whole team's output when they all lean on it. Left unmanaged, an organisation that outsources its ideation to AI is quietly trading distinctiveness for polish.

Why does this matter for a business rather than just for writers?

Because distinctiveness is the whole game. Your advantage is being different in a way your customers value, and if AI is quietly converging your output toward the same middle that everyone else's AI is producing, you end up with efficient sameness: competent, on-trend, and forgettable. The danger is that the homogenisation is invisible where you tend to look. Each individual piece looks good, even better than before, so nothing sets off an alarm. It only becomes visible in aggregate, when you step back and notice it all sounds the same, by which point you have optimised your way into the crowd. This is the same logic behind why your advantage comes from what is distinctive to you, not the tools everyone can buy.

How do I get the lift without the flattening?

By being deliberate about where AI raises the floor and where you protect the peaks.

1. Use AI to raise the floor, not set the ceiling. Let it strengthen first drafts and support weaker contributors, but do not let it originate the distinctive core of your work.
2. Start from the human, then bring AI in. Generate the seed idea before you touch the model, so AI amplifies your distinctiveness rather than replacing it with its average.
3. Protect divergence on purpose. Actively reward the odd, dissenting, off-distribution idea the model would never suggest. That is your originality reserve, and it needs defending precisely because AI makes conformity so easy.
4. Watch aggregate sameness, not just individual quality. Review whether your output is converging over time; judging pieces one by one hides the very effect you need to catch.
5. Keep some AI-free zones. The work where being genuinely different matters most is the work you should not run through an averaging machine.

What does this change for me as a leader?

It changes the question you ask about AI and creativity. Not "does it help", because individually it clearly does, but "what is it doing to our distinctiveness", where the honest answer, left unmanaged, is eroding it. The organisations that win with AI here will not be the ones that adopt it hardest for idea generation. They will be the ones that use it to lift every person's baseline while fiercely protecting the human originality that makes them recognisably themselves.

That is the subtle version of the end of business as usual: when everyone has the same brilliant averaging machine, sameness becomes the default and difference becomes the discipline. Use AI to raise your people, and guard, deliberately, the strangeness that no model will ever hand you. That strangeness is the part customers actually remember.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.