Why Your ChatGPT Ideas All Sound the Same—and How to Fix It

by Claire Brady, EdD

A new study in Nature Human Behaviour, covered by Axios, underscores a crucial point higher ed leaders must internalize: ChatGPT can generate a high volume of creative ideas—but these ideas often lack diversity. Researchers found that over 90% of ChatGPT-generated responses in a brainstorming task shared overlapping concepts. In contrast, human participants using their own ideas and web searches created a much broader range of solutions.

This doesn’t mean ChatGPT isn’t useful. It is—but only when used thoughtfully and in combination with human ideation. And that’s exactly where we, as educational leaders, have an opportunity (and obligation) to step in.

What This Means for Higher Ed

In an era where many institutions are racing to integrate generative AI into everything from classroom teaching to strategic planning, this research is a timely reminder: AI is a tool, not a replacement for human creativity. Too often, we treat AI outputs as polished answers instead of starting points. This mindset can flatten brainstorming, innovation, and even institutional storytelling. If your campus AI strategy assumes that using ChatGPT alone will improve ideation or save time across the board, it’s time to reconsider.

Actionable Guidance for Leaders

1. Reinforce the “Co-Creator” Mindset.

Encourage your teams, faculty, and students to treat ChatGPT as a collaborator, not an oracle. The best ideas emerge when people bring their own perspectives and use AI as a prompt partner—not as a substitute for thinking.

2. Model Better Prompts.

Use "chain of thought prompting" to push ChatGPT toward greater idea diversity.

For example:

“Generate 5 completely different ideas for this initiative—avoid repetition and ensure each is based on a unique concept or angle.”

Build this into faculty and staff training, student AI literacy workshops, and your own leadership practice. AI doesn’t replace divergent thinking—it needs it.

3. Audit for Idea Diversity.

When reviewing AI-assisted projects—marketing campaigns, curriculum redesigns, student success initiatives—ask: Do these ideas feel too similar? Have we pushed past the first, most obvious layer of output? This simple question can guard against echo chamber thinking.

4. Encourage Layered Brainstorming.

Use AI to spark ideas, then convene real humans to expand, challenge, or remix those ideas. Consider a “reverse brainstorm”: have staff review AI-generated ideas and critique what’s missing or where the tool made assumptions.

5. Lead by Example.

When you use ChatGPT or other AI tools, narrate your process. Share not just what the tool gave you, but what you did next—how you pushed the ideas further, discarded the clichés, and infused your own expertise. That’s the behavior worth modeling.

The Bottom Line

AI’s outputs are fast—but they’re rarely final. As higher ed leaders, our role is to shape the conditions for creative, thoughtful, human-centered innovation—with AI as a strategic co-pilot, not the driver.

The future isn’t “AI vs. human creativity.” It’s “AI + human creativity”—when we know how to use both wisely.

Read the full article at: https://www.axios.com/2025/06/18/ai-chatgpt-research-creativity-brainstorming?utm_source=flipboard&utm_content=user%2FAxiosNews