Teacher-provided correct answers and pre-prompting reduce erroneous LLM output and hallucinations compared with open chatbot use¶

Claim · All claims
Evidence · 1 study · design r1 · q1
Currency · newest study 2024 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q1 i? The authors report that having the teacher supply the correct solution and pre-prompt the AI strongly reduces erroneous output and hallucinations, though this is an asserted design rationale rather than a measured comparison. → Steinert 2024

Evidence¶

Steinert 2024¶

Steinert, S., Avila, K. E., Ruzika, S., Kuhn, J., & Küchemann, S. (2024). Harnessing large language models to develop research-based learning assistants for formative feedback. Smart Learning Environments. https://doi.org/10.1186/s40561-024-00354-1

q1 · i? · design · r1

Authors' interpretation in the discussion of LEAP's design, contrasting it with chatbot-style use where students prompt the LLM themselves. No data or error-rate measurements are reported to support the reduction claim.

"it strongly reduces the number of erroneous output and hallucinations because the teacher provides the correct solution to the task, and it does not allow the student to get side tracked as it kindly guides the student back to the task."

Discussion¶

Related Claims¶

• More complex prompts tend to elicit superior results with advanced LLM versions (authors' observation) — related
• Gen-AI hallucinates sources and values and produces convincing but incorrect explanations, risking student misinformation — related
• Providing the correct solution in the prompt improves LLM alignment with human-authored distractors, confirming correct-solution recovery as a primary bottleneck — related
• Reviewed literature reports risks of overreliance, AI errors in complex tasks, and students' uncritical acceptance of AI-generated outputs — related
• Students perceive LLM chatbots as useful for learning but requiring critical use, reporting hallucinations, overconfidence, and dependency risks — related
• The most frequently cited ethical concerns about GenAI in research were hallucinated facts, fabricated references, and misleading content — related