Mitigation strategies for GenAI challenges in PBL (validation, AI-free periods, equity monitoring)¶

Strategy · All strategies
Evidence · 2 claims (2 for) · 2 studies (1 review, 1 theoretical), q1–q2 · 0 of 2 report an effect size · 2 claims rest on one study

Description¶

The article's recommended responses to the risks of GenAI in PBL: rigorous validation of outputs, retrieval-augmented generation anchored in verified literature, explicit training in critical evaluation, defined AI-use boundaries, AI-free problem-solving periods, assessments that evaluate solutions beyond simple adoption, inclusive prompting and multilingual interfaces, and institutional policies for privacy and integrity. For hallucinations, the article states that "Mitigation strategies include implementing rigorous validation protocols, integrating retrieval-augmented generation methods that reference verified biomedical literature, and explicitly training students to evaluate AI outputs critically."

Design Implications¶

Context¶

Requirements¶

• Institutional policies governing AI usage in educational settings
• Rater training with reliability checks to prevent uncritical acceptance of AI-mediated work

Constraints¶

• GenAI models may produce hallucinations, reflect training-data biases, or obscure provenance, and often display reduced effectiveness for non-native English speakers

Target Learners¶

• biomedical engineering students in AI-enhanced courses
• non-native English speakers requiring multilingual interfaces

Target Learning Goals¶

• critical evaluation of AI outputs
• deep conceptual understanding independent of AI assistance
• equitable benefit from AI-enhanced education

Claims¶

• Retrieval-augmented generation reduces but does not eliminate hallucinations in AI-assisted verification [+W]
• LLM responses that are incomplete or incorrect prompt students to question, verify and improve the information, fostering critical thinking through cycles of validation and correction [+W]

Related Strategies¶

• Replace detector-centred assessment with a four-stage process-oriented model: declared starting point, versioned drafting, process artefacts, and targeted oral verification

Examples¶

-

Key Sources¶

• Micky C. Nnamdi, J. Ben Tamo, Benoit Marteau, Wenqi Shi, May D. Wang. (2026). Advancing Problem-Based Learning in Biomedical Engineering in the Era of Generative AI. arXiv. https://arxiv.org/abs/2503.16558