Leading LLMs hallucinated legal authorities in 58%-88% of tested cases while failing to recognise incorrect legal assumptions in prompts¶

Claim · All claims
Evidence · 1 study · theoretical r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? In the legal domain, leading LLMs hallucinated legal authorities in 58%-88% of tested cases and failed to recognise incorrect legal assumptions in users' prompts. → Łucja Biel 2026

Evidence¶

Łucja Biel 2026¶

Łucja Biel. (2026). AI Literacy for Legal Translation: Developing Digital Resilience. University of Warsaw. https://arxiv.org/abs/2608.04641 doi:10.48550/arxiv.2608.04641

q2 · i? · theoretical · r2

The chapter reports, citing Dahl et al. (2024), that leading LLMs "hallucinated legal authorities in 58% -88% of tested cases"; the reviewed study's design is not described in the chapter, so this is a second-hand citation.

"Dahl et al. (2024) found that leading LLMs hallucinated legal authorities in 58% -88% of tested cases while failing to recognise incorrect legal assumptions in users’ prompts."

Discussion¶

Related Claims¶

• LLM-assisted inductive qualitative coding carries risks of superficial themes, broad or redundant codes, and hallucinated interpretations, so LLMs should augment rather than replace human researchers — related
• Existing information retrieval systems fail to support users' constructive process in complex tasks — related
• Zero-shot LLM errors follow four recurring patterns: over-interpretation, failure to detect relevant information, hallucination, and failure to generate a response
• Gen-AI hallucinates sources and values and produces convincing but incorrect explanations, risking student misinformation
• Of 30 verified hallucinated references, 13 appeared entirely fabricated and 17 were hybrid references combining a real title with fabricated authorship or metadata
• 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
• Reviewed literature reports risks of overreliance, AI errors in complex tasks, and students' uncritical acceptance of AI-generated outputs