Document Type
Research Brief
DOI
10.5703/1288284318572
Start Date
10-11-2025 12:00 AM
Description
This comparative study investigates the different emphases of the four mainstream GenAI tools when adopted to design interdisciplinary learning projects for the 9-12 STEM education. It addresses a gap in understanding the specialties of GenAI tools when used to support instructional design in STEM education (Zekaj, 2023). The researchers gave the same prompts for designing a project-based learning session to the GenAI tools: ChatGPT 5, Gemini 2.5 Pro, Copilot, and Claude Sonnet 4.5. The AI-generated outputs were evaluated using the Expert Judgement framework (Dick et al., 2015). Each GenAI revised its own design based on the expert feedback, and the second-round results were inter-evaluated based on the inputs from the first round. A comparison of the final results from the second round of GenAIs was conducted to identify the different emphases of the GenAI tools in designing PBLs. This study contributes to understanding how iterative GenAI-assisted instructional design can refine PBL designs and informs instructional designers how to leverage the advantages of GenAI in designing interdisciplinary PBL for K-12 STEM education. As a DBR, the design revision decision-making process can provide insights for future design considerations.
A Comparative DBR Evaluation for the Four Mainstream GenAI Tools in Designing PBL for 9-12 STEM Education
This comparative study investigates the different emphases of the four mainstream GenAI tools when adopted to design interdisciplinary learning projects for the 9-12 STEM education. It addresses a gap in understanding the specialties of GenAI tools when used to support instructional design in STEM education (Zekaj, 2023). The researchers gave the same prompts for designing a project-based learning session to the GenAI tools: ChatGPT 5, Gemini 2.5 Pro, Copilot, and Claude Sonnet 4.5. The AI-generated outputs were evaluated using the Expert Judgement framework (Dick et al., 2015). Each GenAI revised its own design based on the expert feedback, and the second-round results were inter-evaluated based on the inputs from the first round. A comparison of the final results from the second round of GenAIs was conducted to identify the different emphases of the GenAI tools in designing PBLs. This study contributes to understanding how iterative GenAI-assisted instructional design can refine PBL designs and informs instructional designers how to leverage the advantages of GenAI in designing interdisciplinary PBL for K-12 STEM education. As a DBR, the design revision decision-making process can provide insights for future design considerations.