What Makes a Good Prompt? How Students Use ChatGPT to Solve Creative Problems

What is it about?

In this study, we wanted to understand how university students actually use ChatGPT when they face a creative problem. Many of us now rely on AI tools to help us brainstorm ideas, yet we still know very little about which kinds of prompts help students think more creatively. To explore this, we analysed the full conversations of 77 students who used ChatGPT to generate new ideas for improving a children’s toy as part of a creative problem-solving task. We looked closely at the wording of the prompts students used and how these prompts shaped the quality of the ideas they eventually produced. What became clear is that not all prompting strategies are equal. Students who performed well tended to treat ChatGPT like a collaborative partner. They gave clear instructions, included useful details about the task, and offered specific feedback when the AI’s suggestions were not original or practical enough. This created an active, back-and-forth dialogue in which students and ChatGPT worked together and refined the ideas step by step. Students who produced weaker ideas usually used ChatGPT in a very different way. Instead of engaging in a conversation, they asked simple factual questions about rabbits, toys, or children’s preferences. This style of interaction resembled a basic search query and did not help them generate original or useful ideas. As a result, their final solutions tended to be predictable and similar to common answers found online. Overall, our study shows that creativity with AI does not happen automatically. Good prompting requires giving the AI clear expectations, adding context, and responding to its suggestions with thoughtful feedback. These strategies help people guide the AI toward more innovative and meaningful outcomes. Our findings offer practical insights that can help students, teachers, and professionals use AI tools more effectively for creative work.

Featured Image

Why is it important?

This work is important because generative AI tools are now widely used in education, yet there is little evidence on how people should interact with them to achieve high-quality results. Many users assume that simply asking for ideas is enough, but our findings show that effective collaboration with AI requires active guidance and thoughtful prompting. At a time when universities and workplaces are rapidly adopting AI, our study provides timely evidence that can inform teaching practices, digital literacy training, and responsible AI use. By identifying the specific strategies that support creative thinking, we can help people use AI in ways that enhance originality rather than limit it.