Providing LLM pre-labels can decrease human annotation accuracy when the model is incorrect, due to automation bias¶

Claim · All claims
Evidence · 1 study · causal r? · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Demonstrated research shows that providing LLM pre-labels can decrease human accuracy when the model is incorrect, as annotators may uncritically adopt the AI's suggestion. → Thomas 2026

Evidence¶

Thomas 2026¶

Thomas, D.R., Borchers, C., Vanacore, K.P., Koedinger, K.R., & Kizilcec, R.F. (2026). Modernizing Ground Truth: Four Shifts Toward Improving Reliability and Validity in AI in Education. https://arxiv.org/abs/2603.29141 doi:10.48550/arxiv.2603.29141

q2 · i? · causal · r?

The article attributes to prior work the demonstration that LLM pre-labels "can actually decrease human accuracy when the model is incorrect" because annotators uncritically adopt the AI's suggestion. The study's design is not described in this article.

"[50] demonstrated that providing LLM pre-labels can actually decrease human accuracy when the model is incorrect, as annotators may uncritically adopt the AI's suggestion"

Discussion¶

Related Claims¶

• Reviewed literature reports risks of overreliance, AI errors in complex tasks, and students' uncritical acceptance of AI-generated outputs — a broader claim this one bears on
• A pilot human annotation surfaces a framing discrepancy: a NEET aspirant labeled all 10 sampled questions as Recall, including the five the LLM tagged as Elimination/Negation — related
• Automated AI assessment systems reduce grading time and improve evaluation consistency in nursing education — related
• Warning labels alone did not improve performance or reduce reliance on GenAI advice, and students sometimes adopted incorrect AI answers — related
• Automation bias toward fluent LLM output surrenders learners' epistemic agency — related