A new study from Georgetown University warns that generative AI can produce text with richer vocabulary but fewer original ideas, challenging the assumption that AI creativity matches human creativity. For creatives who rely on idea generation, the research suggests starting with pen and paper before turning to AI.
Adam Green, a neuroscientist and psychology professor who studies creativity, led the research. He said AI can be creative, but the danger lies in treating it as a substitute for human thought. "I have no problem saying that AI can be creative," Green said. "What I think we have to be careful of is to assume that it's a replacement for human creativity."
How the brain generates originality
Human creativity operates on semantic networks-personal webs of experience stored in the brain. When you learn something, information is wired along these networks. During a creative act, the brain searches this unique map to combine old ideas into new ones. Because no two life experiences are identical, individual brains produce ideas that differ from everyone else's.
AI, by contrast, relies on a single shared system. "When you ask AI to do something creative, it will do something creative," Green said. "It can write you a poem. It can try to come up with an unusual use for a paperclip. But the ideas it produces, even though they may be good ideas, end up pretty similar to each other. What's lost is the human diversity of ideas."
What the college essay data reveals
Green and his colleagues analyzed over 370,000 college admissions essays written before and after ChatGPT's launch in 2022. Essays that used AI assistance showed a greater diversity of words. However, they contained fewer original ideas.
The study found that human readers often mistake varied vocabulary for true creativity. Outside experts rated the AI-assisted essays as more creative, even when the ideas were not more original. This decoupling of word choice from idea novelty is a new phenomenon, the researchers say, because historically, a richer vocabulary signaled deeper thinking.
The long-term impact worries Green. He compared AI's effect on cognition to how GPS navigation eroded spatial reasoning. "It used to be that people could navigate on their own and had a lot of spatial cognitive abilities. GPS has totally changed that, and people cannot navigate without their phone or GPS. As people have lost this spatial ability, it is very likely they're losing more general cognitive abilities," he said.
How to use AI as a creative partner, not a replacement
Green does not reject AI. He recommends creatives always start with their own ideas. Once a concept exists, AI can sharpen it, analyze it for flaws, or assist with editing. In his lab, researchers use AI to process data, but the initial questions and parameters come from human minds.
"AI can help me with the less creative part of the thinking, and it might even free me up to do more creative thinking because I don't have to spend time doing that more mindless stuff," Green said. For creatives, understanding how to use AI as a complement to human idea generation is crucial, and resources like AI for Creatives can help navigate this balance.
The future of human creativity
Green's research also explores how electrical and magnetic stimulation can boost the brain's creative networks. His lab is testing transcranial electric stimulation devices that look like headsets. Early findings suggest repeated stimulation over weeks may rewire neural pathways for creative thinking, offering benefits that last far longer than a single session.
"What AI is doing in a way that nothing else ever has is asking us to trade that process in favor of product," Green said. "All of the benefits that come with creative engagement are process-related." He adds that the value of a truly novel idea far outweighs a polished but derivative one. "If I can generate something good and truly different, that has much more value than something good but the same."
Why this matters for creatives
This research underscores that the creative process-the messy, personal act of generating ideas-is where the real value lies. Relying on AI to shortcut that process may produce faster output, but it risks flattening the distinct perspectives that make creative work stand out. For writers, designers, and strategists, the takeaway is practical: begin with your own thoughts, then use AI as a sharpening tool, not a starting point. Protecting the habit of original ideation now could be the difference between solving tomorrow's complex problems and recycling yesterday's answers.
Your membership also unlocks:
