05/29/2026
Legal tech used to be something most lawyers could safely ignore. 👋
You needed an email. You needed PDFs. You needed to survive the firm’s document-management system. But the machinery underneath? That could mostly stay with IT. 🖥️
AI now touches legal research, contract review, client intake, litigation strategy, legal writing, court filings, and more. Lawyers do not need to become engineers, but they do need a working vocabulary. 📑
Here are the AI concepts every lawyer should know by now:
1️⃣ Artificial intelligence is the broad category. It covers tools that do tasks we associate with human intelligence.
2️⃣ When people talk about “using AI,” they usually mean they are using a model. A model is the trained system behind the tool.
3️⃣ A large language model, or LLM, is the kind of model behind tools like ChatGPT, Claude, or Gemini.
4️⃣ Training data is the material used to build the model. It helps the model learn language, concepts, relationships, patterns, and forms.
5️⃣ Context is the material available to the model in the current task: your prompt, pasted text, uploaded documents, or chat history.
6️⃣ Models process text in pieces called tokens. A token may be a word, part of a word, punctuation, or another chunk of text.
7️⃣ A prompt is the instruction you give the tool.
8️⃣ Hallucination is when AI produces false information as if it were true.
9️⃣ A RAG (retrieval-augmented generation) system first retrieves relevant information from a set of documents or database. Then the model uses that retrieved material to generate an answer.
🔟 An AI agent is a system that can take steps toward a goal. That may mean searching files, drafting a response, updating a spreadsheet, creating a task, checking a database, or using other tools.
Lawyers do not need to know every technical layer. But they do need to understand the concepts that shape the risk. ☑️
-