Show GN: Lemini — 두 가지 모드로 동작하는 법 자문 챗봇

Lemini distinguishes itself by recognizing that legal queries fall into two distinct categories—open-ended advice and structured document review—each requiring fundamentally different reasoning chains rather than a single RAG approach. Its citation verification loop directly addresses the hallucination problem endemic to legal LLMs by cross-referencing claims against actual Korean statutes, precedents, and regulatory documents, making auditability a core feature rather than an afterthought. This signals an important shift in legal AI adoption: users prioritize evidence-grounded answers with verifiable sources over conversational polish, particularly in high-stakes domains where accuracy determines legal risk.