AI Beats Typical Students but Loses to Top Papers in Model UN

📘 Context — Why Model UN Papers Matter

Model United Nations position papers demand extensive background research and concise policy articulation on international issues. This study assesses how well ChatGPT can produce those papers compared with student-written submissions.

📚 What Was Compared

• Student-written Model UN position papers
• ChatGPT-generated Model UN position papers

Both paper types were evaluated on their quality in the Model UN context, where clear, researched, and policy-focused writing is central.

🧾 Key Findings

• AI-generated papers received higher evaluations overall than the typical student submissions.
• Award-winning student papers outperformed ChatGPT-generated papers, demonstrating that top human work still exceeds AI in this task.
• The quality of AI-generated writing depends on human input, highlighting that AI output varies with how humans interact with and shape it.

🎯 Why It Matters

These results show a dual reality for educators and assessors: generative AI can produce strong, evaluable position papers that often surpass average student work, but the best student writing remains superior. The dependence of AI quality on human input underscores the need for a human-centered approach to integrating AI into classrooms, assessments, and pedagogy.