Engage students in active learning assignments using generative AI with structured reflection and critique of AI outputs¶

Strategy · All strategies
Evidence · 2 claims (1 mixed, 1 against) · 3 studies (3 associational), q1–q3 · 0 of 3 report an effect size · 1 claim rests on one study

Description¶

The review recommends that educators craft assignments incorporating generative AI that require students to critique AI-generated text and reflect on what the AI did well or poorly, rather than banning the tools. Across practices, "a common thread is structured reflection": debrief questions or discussions where students articulate the AI's performance and how they arrived at their final answer. This treats AI as a partner to question and improve upon, illuminating its affordances and limitations and improving students' evaluative skills.

Design Implications¶

Context¶

Requirements¶

• Instructors must guide use and include debrief questions or discussions where students articulate what the AI did well or poorly

Constraints¶

• The review warns that without guidance students may over-rely on AI, so instructional design should phase AI support or require students to attempt tasks themselves before consulting AI

Target Learners¶

• higher education students across disciplines, including first-year writing and STEM students

Target Learning Goals¶

• AI literacy
• critical evaluation of AI outputs
• feedback literacy

Claims¶

• Text-based reflective use from Copilot Critical Engagement responses does not predict productive engagement, with limited coder reliability and a TA ceiling effect (exploratory null) [-W]
• Students frequently accepted misleading AI recommendations (overreliance) while rarely rejecting correct ones (underreliance) [~W]

Related Strategies¶

• Resistance Reading and Counter-Narratives
• Require students to justify how they used AI output: what they accepted, rejected, revised, or verified, and why
• Support learners to evaluate AI outputs by verifying sources, reporting errors, and overriding recommendations
• Computing instructors should approach GenAI slide generation as programmers: work iteratively with coding assistants and text-based slide toolchains
• Use shared metacognition as the routine practice for educators and learners to critically evaluate AI-generated outputs
• Embed AI as a learning partner and redesign assessments for AI-rich contexts
• Use structured exercises requiring students to compare, critique, and justify against AI-generated code
• Address generative AI in higher education through assessment redesign and proportional reflective practices rather than surveillance or punitive controls
• Design PjBL projects so students must evaluate, critique, and improve AI-generated outputs rather than prohibiting AI use
• Incorporate collaborative learning activities within AI literacy teaching, attending to activity modality and integration with other activities and assessments
• Embed AI in guided accept/reject/justify workflows, teach digital critical thinking explicitly, and integrate institutional authorship standards
• Iteratively add prompts requiring comparison of AI outputs with psychological theories and structured reflective prompts to counter overreliance on AI
• Treat AI literacy as educational support by teaching students to question AI responses, attempt problems first, and avoid substituting AI output for engagement
• Embed ethical inquiry into AI-supported instruction by having learners compare AI-generated texts with peer-reviewed sources
• Require students to justify AI-assisted solutions and compare AI-generated with self-developed code

Examples¶

-

Key Sources¶

• Jacob, S., Miceli, H., & Schneider, H. (2026, May). Landscape of emerging technologies in higher education: A review. Digital Promise & The American Association of Colleges & Universities. https://doi.org/10.51388/20.500.12265/293