AI glossary · Basic · Language models · Legal & professional services
Legal AI
Legal AI is technology that uses AI to help find, read, summarize and draft legal documents such as case law, statutes and contracts. Rather than replacing lawyers or legal teams, it is mainly used as a tool to reduce repetitive document work.
About this term
- Level
- Basic · Technical terms you often meet in AI news
- Category
- Industry AI
- Fields
- Language models
- Industries
- Legal & professional services
- Added
- 2026-10-10
How it works
Most systems work through RAG, attaching legal document search to an LLM (large language model). When a question comes in, the system first finds relevant passages in case law and statute databases or the company's internal contracts, then generates an answer or summary based on that content. In contract review, it splits the contract into individual clauses, compares them with standard wording, and flags clauses that are unfavorable or missing. When preparing for litigation, it is also used to pick out relevant documents from vast amounts of evidence.
Why it matters
Much of legal work consists of carefully reading and comparing large numbers of documents, which takes a long time and costs a lot. When AI handles first-pass review and drafting, lawyers can focus more on work that requires a person, such as judgment, strategy and client consultation. There is also hope that if the cost of legal services falls, individuals and small businesses will be able to get basic legal information more easily.
Things to know
The biggest risk is hallucination. Because the system can plausibly cite cases or provisions that do not actually exist, a procedure for checking cited sources directly against the originals is essential. Client information is sensitive data subject to confidentiality obligations, so you must check whether it may be uploaded to an external AI service and whether it will be used for training. In addition, many countries have rules restricting non-lawyers or services from giving legal advice, so it is emphasized that responsibility for AI output ultimately lies with the professional who reviewed and used it.
Checklist before adopting legal AI
Accuracy
- Have cited cases and provisions been checked against the originals?
- Does it show the source documents behind the answer?
Security · responsibility
- May client information be sent outside?
- Is input material excluded from training?
- Is a final reviewer designated?
Example
Consider a corporate legal team that needs to renew hundreds of contracts with business partners. Legal AI extracts the liability cap, termination conditions and confidentiality clauses from each contract, organizes them into a table, and flags clauses that differ from the company standard. The lawyer in charge reads only the flagged clauses closely and decides how to revise them. This division of labor, where AI drafts and people make the judgment, is currently the most common use.