Think first, ChatGPT later: a three-phase approach for guided human–AI collaboration in creative problem-solving¶

Strategy · All strategies
Evidence · 4 claims (4 for) · 1 study (1 causal), q3 · 1 of 1 report an effect size · 4 claims rest on one study

Description¶

A structured approach guiding learners to interact collaboratively with ChatGPT during creative problem-solving. In Phase 1, learners independently generate their own initial ideas. In Phase 2, they interact collaboratively with ChatGPT to "improve, develop, and evaluate" those ideas using sample prompts that scaffold creative and metacognitive processes. In Phase 3, they independently propose a single best solution. The approach was tested against free ChatGPT use and independent work.

Design Implications¶

Context¶

Requirements¶

• Learners must have access to ChatGPT during Phase 2
• Sample prompts or equivalent scaffolding for collaborative interaction must be provided
• Learners must be instructed to treat ChatGPT as a brainstorming and feedback partner rather than a provider of final answers

Constraints¶

• Tested with university students on creative product improvement and invention tasks
• The assisted-task performance of the regulated-AI group did not exceed independent work, so the approach trades immediate performance for learning gains

Target Learners¶

• university students

Target Learning Goals¶

• independent creative problem-solving
• original and useful idea generation
• metacognitive regulation during creative tasks

Claims¶

• Collaborative prompt use mediated the regulated-AI group's advantage over the general-AI group in independent originality [+W]
• The general-AI group most often dictated ChatGPT to directly generate solutions while the regulated-AI group more frequently collaborated to improve self-generated ideas [+W]
• The regulated 'think first, ChatGPT later' group outperformed both other groups in independent originality on the unassisted task despite no assisted-task advantage [+W]
• The regulated-AI group sustained originality across tasks while both other groups declined from the easier to the more complex task [+W]

Related Strategies¶

• Combine GAI tools with traditional methods in a hybrid approach to balance efficient collaboration and critical evaluation
• Use ChatGPT as a tool for supporting students’ learning and promoting problem-solving skills

Examples¶

-

Key Sources¶

• Wong, S. S. H., & Qiu, S. X. (2026). Think First, ChatGPT Later: Guiding Human–AI Collaboration for Learning Gains in Independent Human Creativity. Educational Psychology Review. https://doi.org/10.1007/s10648-026-10118-7