AI & Generative AI Glossary: 30+ Terms Explained (2026 Edition)

AI and Generative AI terminology changes fast — often faster than most glossaries can keep up with. This page is a focused, plain-English reference for the terms you'll actually run into while learning GenAI, building with AI tools, or reading about AI in the news. No fluff, no filler — just clear definitions you can bookmark and come back to.

New to this space? DizitalAdda's Diploma in GenAI & Prompt Engineering covers every term on this page as a hands-on module, not just a definition.

Core AI Concepts

• Artificial Intelligence (AI) The broad field of building systems that can perform tasks — recognizing patterns, understanding language, making decisions — that would normally require human intelligence.
• Machine Learning (ML) A subset of AI where systems improve their performance on a task by learning patterns from data, rather than being explicitly programmed with fixed rules.
• Deep Learning A type of machine learning that uses multi-layered neural networks to learn from large amounts of data, forming the foundation of most modern AI systems, including LLMs.
• Neural Network A computing system loosely inspired by the human brain's structure, made up of layers of interconnected nodes that process and learn from data.
• Multimodal AI AI systems that can understand and generate more than one type of content — for example, processing both text and images within a single model.
• Token The basic unit of text an AI language model processes — roughly a word or part of a word — used to measure both input length and processing cost.

Want the fundamentals in a structured course? See DizitalAdda's Machine Learning & Cloud Computing track.

Generative AI & LLMs

• Generative AI (GenAI) AI systems that create new content — text, images, audio, code — rather than just analyzing or classifying existing data.
• Large Language Model (LLM) An AI model trained on massive amounts of text data to understand and generate human-like language; ChatGPT, Claude, and Gemini are all powered by LLMs.
• Context Window The maximum amount of text (measured in tokens) an AI model can consider at once when generating a response — anything beyond this limit is not "seen" by the model.
• Hallucination When an AI model generates an answer that sounds confident and plausible but is factually incorrect or fabricated. IBM's explainer on AI hallucinations is a solid non-vendor-specific primer on why this happens.
• Fine-tuning The process of further training an existing AI model on a specific, narrower dataset to adjust its behavior, tone, or knowledge for a particular use case.
• Training Data The large dataset used to teach an AI model patterns and knowledge before it's made available for use.
• Model Weights The internal numerical parameters an AI model learns during training, which determine how it processes input and generates output.

RAG & Retrieval

RAG (Retrieval-Augmented Generation) A technique where an AI model retrieves relevant information from an external source — like a document database — before generating an answer, reducing hallucination and allowing the model to use current or private data instead of relying only on what it memorized during training. Read DizitalAdda's full walkthrough: What is RAG? Explained for Beginners →

• Vector Database A specialized database that stores information as numerical representations (embeddings) so it can be searched based on meaning and similarity, rather than exact keyword matches — a core component of RAG systems.
• Embedding A numerical representation of text (or images, audio, etc.) that captures its meaning, allowing AI systems to measure how similar two pieces of content are.
• Semantic Search A search method that matches results based on meaning and intent rather than exact keyword overlap — made possible by embeddings and vector databases.
• Chunking The process of breaking a large document into smaller sections before converting it into embeddings, so a RAG system can retrieve only the most relevant piece rather than an entire document.

AI Agents

• Agentic AI / AI Agent An AI system that can take a goal, break it into steps, and act autonomously — using tools, making decisions, and adjusting its approach — rather than simply responding to a single prompt.
• Chatbot A conversational AI system designed to respond to user messages, typically limited to answering or generating text within a single exchange, without autonomously taking multi-step actions.
• Tool Use / Function Calling An AI model's ability to call external tools, APIs, or functions (like a calculator, search engine, or database) as part of generating a response, rather than relying purely on its own internal knowledge. OpenAI's function calling guide is a widely referenced technical overview of the pattern.
• Multi-Agent System An AI setup where multiple specialized AI agents work together, each handling a different part of a task, and coordinating to reach a shared goal.
• Autonomous Workflow A sequence of tasks an AI agent completes on its own, with minimal human intervention, often by chaining together multiple steps or tools.

This is one of the fastest-growing areas of AI hiring. DizitalAdda's Diploma in GenAI & Prompt Engineering includes dedicated modules on building multi-agent systems.

Prompting & Model Behavior

• Prompt Engineering The practice of carefully wording instructions given to an AI model to get more accurate, useful, or consistent outputs.
• System Prompt A set of instructions given to an AI model before a conversation begins, defining its behavior, role, or constraints for that session.
• Zero-shot Prompting Asking an AI model to complete a task without giving it any prior examples, relying entirely on its existing training.
• Few-shot Prompting Giving an AI model a small number of examples within the prompt itself to guide it toward the desired output format or style.
• Chain-of-Thought Prompting A prompting technique that encourages an AI model to reason through a problem step-by-step before giving a final answer, often improving accuracy on complex tasks. This approach was first formalized in a widely cited Google Research paper.
• Temperature (AI) A setting that controls how random or predictable an AI model's output is — lower temperature produces more focused, consistent answers, while higher temperature produces more varied or creative ones.

Tools & Frameworks

• LangChain A popular open-source framework used to connect language models with external data sources, tools, and memory, commonly used to build RAG systems and AI agents. See the official LangChain documentation for current guides.
• LangGraph A framework built on top of LangChain for creating more complex, stateful AI agent workflows, where the sequence of steps can branch or loop rather than running in a straight line. Reference: LangGraph documentation.
• API (Application Programming Interface) A set of rules that allows different software systems to communicate with and use each other's functionality — the standard way most AI models are integrated into apps.
• Model Context Protocol (MCP) A standard that allows AI models to connect with external tools and data sources in a consistent way, making it easier to build AI systems that can interact with multiple apps and services. See the Model Context Protocol specification for technical details. 

DizitalAdda's Advanced Certification in GenAI & Prompt Engineering covers LangChain, LangGraph, RAG pipelines, and MCP-based tool integration through live, job-ready projects.

Ready to Build with These Concepts, Not Just Know Them?

Reading a glossary tells you what RAG or an AI agent is. Building one — end to end, with a real vector database and a working multi-agent workflow — is what actually shows up on a portfolio. That's the gap DizitalAdda, Delhi-NCR's AI, data science, and digital marketing training institute, is built to close. Since 2009, DizitalAdda has trained 25,000+ students through small batches, live projects, and mentors who work with these tools professionally.

If you're just starting out, the Diploma in GenAI & Prompt Engineering (12 months) builds from fundamentals through ReAct, LangGraph, and multi-agent systems. If you already know the basics and want to move faster, the Advanced Certification in GenAI & Prompt Engineering (6 months) gets you building chatbots, RAG pipelines, and GenAI apps hands-on. 

Explore all DizitalAdda AI courses → or book a free demo class to see the curriculum in action before you enrol.

This glossary focuses specifically on AI and Generative AI terminology. For broader digital marketing, SEO, and data analytics terms, see our Digital Marketing Glossary → — link once that post is live.

This page is updated regularly as new AI terminology enters common use. If there's a term you'd like explained that isn't here yet, let us know.