REZA LEGALrezalegal.nl ↗
Legal Tech & AI Tools — Bronnen

L027

Stanford Legal RAG Hallucination Study — Evaluatie van AI Legal Research Tools

Onderzoek-LiteratuurOriënterendUSSecundair
Rechtsgebied
AI-recht / ICT-recht
Rechtsvraag
Hoe betrouwbaar zijn AI-aangedreven legal research tools (LexisNexis, Thomson Reuters)?
Kernantwoord
Ondanks providerclaims dat RAG 'hallucinatie-vrij' is, hallucineren Lexis+ AI en Westlaw AI-tools nog altijd 17-33% van de tijd — lager dan generieke chatbots, maar substantieel en risicovol in juridische context.
Praktische betekenis voor REZA Legal
Het belangrijkste bewijsstuk waarom AI legal research nooit blind kan worden vertrouwd: menselijke verificatie is een kwestie van professionele zorgplicht.

Audit & Herkomst

Bindendheid
Oriënterend
Bron
Stanford RegLab — Legal RAG Hallucination Study
Brongehalte
Secundair
Laatst gecontroleerd
2026-08-13