A massive new experiment compared leading AI chatbots against more than 100,000 people on creativity tests — and found something that might surprise (or scare) you: today’s top AI can beat the average human on certain “idea generation” tasks. On measures of divergent thinking — the ability to produce unusual, unrelated concepts — models like GPT-4 often generated answers that scored higher than typical human responses.
But here’s the twist: the most creative humans still win, decisively. When researchers looked at the top half of participants — and especially the top 10% — human imagination pulled ahead of every AI system tested. The gap widened even more on richer work like poetry, story writing, and original narrative crafting, where human voice and taste still show up in ways machines struggle to match.
The takeaway isn’t “AI replaces creators.” It’s “AI raises the baseline.” Creativity becomes more like a power tool: capable of generating strong raw material fast, but still needing a human to steer it, refine it, and inject the kind of originality that comes from lived experience, risk-taking, and genuine perspective.
