Scaffolded LLM-collaboration assignment cycle: draft, LLM critique, then rebut or incorporate¶

Strategy · All strategies
Evidence · no claims cited

Description¶

A repeatable classroom assignment cycle in which students first write an argumentative essay defending a position, then ask an LLM chatbot of their choice to critique the argument as thoroughly as possible, then respond by either rebutting the LLM's arguments or updating their original argument. The article reports this cycle was assigned five times in one semester, with prompt-engineering quality graded and a Guidelines for Prompt Engineering document provided at the start.

Design Implications¶

Context¶

Requirements¶

• Students need guidance on effective prompting, such as the two-page Guidelines for Prompt Engineering document the teaching team provided; grading must reward quality of response to the LLM's critique

Constraints¶

• The article reports the study was exploratory and could not isolate the causal effects of the intervention; assignments were graded by a teaching assistant who was not blind to the timing of the assignment

Target Learners¶

• college students in introductory social science courses

Target Learning Goals¶

• argumentative writing
• responding to critical feedback
• prompt engineering

Affordances¶

• Llm Collaborative Learning Partners Framework

Related Strategies¶

• Use Structured Critique Protocols
• Implement AI-peer integrated feedback as a standardized two-phase revision cycle with rubric-aligned prompting and instructor monitoring
• Scenario B: constructing, critiquing, and reframing AI-mediated arguments

Examples¶

-

Key Sources¶

• Daniel M. Oppenheimer, Trent N. Cash, Allison E. Connell Pensky. (2025). You've Got AI Friend in Me: LLMs as Collaborative Learning Partners. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-025-00521-6