This glossary breaks down the most important terms in artificial intelligence and legal technology in alphabetical order. Each definition explains complex concepts in plain English so lawyers, compliance professionals, and technologists can understand how AI is reshaping the legal profession. Most include concrete examples to assist understanding.
AI Governance: The rules and oversight that keep AI use fair, transparent, and compliant with laws and ethical standards.
AI Hallucination: When an AI gives a confident answer that sounds real but is actually false or completely made up.
Algorithmic Bias: When an AI system makes unfair decisions because it learned from data that already contained bias or inequality.
Algorithmic Transparency: The ability to see and understand how an AI system makes its decisions, helping ensure fairness and accountability.
Data Provenance: The record of where data came from and how it was collected or used, ensuring accuracy and legal compliance.
Explainable AI (XAI): AI that can clearly show how and why it made a decision, so people can understand and trust its results.
Fine-Tuning: Teaching an AI extra details with new data so it performs better on a specific task.
Large Language Model (LLM): An advanced AI that learns from huge amounts of text to understand and generate human-like language.
Machine Learning: A way for computers to learn from data and make better decisions over time without being given exact instructions.
Model Drift: When an AI system becomes less accurate over time because the real world changes and its training data no longer matches.
Natural Language Processing (NLP): The technology that helps computers understand and use human language in text or speech.
Neural Network: A computer system that learns and makes decisions by imitating how the human brain processes information.
Overfitting: When an AI performs well on what it already knows but fails when faced with new or different information.
Prompt Engineering: Prompt engineering is the skill of writing clear, detailed instructions that help AI tools produce accurate and useful responses.
Retrieval-Augmented Generation (RAG): A technique where AI pulls in relevant documents first and then uses them to generate accurate legal answers.
Reinforcement Learning: A way for AI to learn by trying, making mistakes, and improving based on rewards and feedback.
Tokenization: The process of breaking text into small pieces called tokens so AI systems can read and understand it.
Training Data: The information used to teach an AI system how to make predictions or decisions.