Skip to main content

AI Terminology Glossary

A beginner-friendly reference for common AI terms and concepts. Use this glossary whenever you encounter unfamiliar words in your AI learning journey.

Core AI Concepts​

Artificial Intelligence (AI) Computer systems that can perform tasks typically requiring human intelligence, like understanding language, recognizing images, or making decisions.

Machine Learning (ML) A type of AI where computers learn patterns from data without being explicitly programmed for each specific task. Like teaching a child to recognize cats by showing them many cat photos.

Deep Learning A subset of machine learning that uses neural networks with many layers. Gets its name from these "deep" networks that can learn complex patterns.

Neural Network A computer system inspired by how brain neurons connect. Processes information through interconnected nodes that can learn to recognize patterns.

Algorithm A set of rules or instructions that tells a computer how to solve a problem or complete a task. Like a recipe for cooking, but for computers.

Language AI Terms​

Large Language Model (LLM) AI systems trained on vast amounts of text to understand and generate human language. Examples include GPT-4, Claude, and Gemini.

Natural Language Processing (NLP) The field of AI focused on helping computers understand, interpret, and generate human language.

Prompt The text input you give to an AI system. Your question, request, or instruction that tells the AI what you want it to do.

Prompt Engineering The skill of writing effective prompts to get better results from AI systems. Like learning how to ask questions that get you the answers you need.

Token The basic unit of text that AI systems process. Roughly equivalent to words or parts of words. Most AI systems have limits on how many tokens they can process at once.

Context Window The amount of text an AI system can "remember" and work with at one time. Like the AI's short-term memory capacity.

Fine-tuning The process of training an AI model on specific data to make it better at particular tasks. Like giving a general-purpose AI specialized training for medical or legal work.

Training and Data Terms​

Training Data The text, images, or other information used to teach an AI system. Like textbooks and practice problems used to educate students.

Dataset A organized collection of data used for training AI systems. Could be millions of books, photos, or other examples.

Parameters Numbers inside an AI model that determine how it processes information and generates responses. More parameters generally mean more sophisticated capabilities.

Bias When an AI system reflects unfair preferences or prejudices from its training data. Can lead to discriminatory or inaccurate outputs for certain groups.

Overfitting When an AI model becomes too specialized on its training data and performs poorly on new, unseen information.

Generalization An AI system's ability to apply what it learned during training to new situations it hasn't seen before.

AI Capabilities and Limitations​

Hallucination When an AI system generates information that sounds plausible but is actually false or made up. Like a confident-sounding wrong answer.

Alignment How well an AI system's behavior matches human values and intentions. The goal of making AI helpful, harmless, and honest.

Emergent Behavior Unexpected capabilities that appear in AI systems as they become more sophisticated, even though they weren't explicitly programmed.

Inference The process of an AI system generating a response or making a prediction based on input. What happens when you submit a prompt and get an answer.

Latency The time delay between sending a request to an AI system and receiving a response. Lower latency means faster responses.

Practical AI Terms​

API (Application Programming Interface) A way for different software programs to communicate. Many AI services offer APIs so developers can integrate AI into other applications.

Multimodal AI AI systems that can work with different types of input like text, images, audio, and video, not just one type.

Few-shot Learning An AI's ability to learn new tasks from just a few examples, rather than needing thousands of training instances.

Zero-shot Learning When an AI can perform tasks it wasn't specifically trained on, based on its general knowledge and understanding.

Retrieval-Augmented Generation (RAG) A technique that combines AI language generation with the ability to look up specific information from databases or documents.

AI Safety and Ethics Terms​

Adversarial Examples Inputs deliberately designed to fool AI systems into making mistakes. Used to test AI robustness and security.

Explainable AI (XAI) AI systems designed so humans can understand how they make decisions. Important for trust and accountability.

Red Teaming Testing AI systems by trying to find ways they might fail, produce harmful outputs, or be misused.

Responsible AI The practice of developing and deploying AI systems in ways that are ethical, fair, and beneficial to society.

Business and Application Terms​

AI Agent An AI system designed to take actions and make decisions autonomously to achieve specific goals. Can interact with other systems and tools.

Automation Using AI or other technology to perform tasks that would normally require human intervention.

Digital Twin A virtual model of a real-world system that uses AI to simulate and predict behavior.

Computer Vision AI that can analyze and understand images and videos. Enables features like face recognition and medical image analysis.

Synthetic Data Artificially generated data that mimics real data but doesn't come from actual events or people. Used when real data is scarce or sensitive.

Model Types and Architectures​

Transformer A type of neural network architecture that's particularly good at processing sequential data like text. The foundation for most modern language models.

Generative AI AI systems that create new content like text, images, music, or code rather than just analyzing existing information.

Discriminative AI AI systems that classify or categorize existing information rather than generating new content.

Foundation Model Large AI models trained on broad data that can be adapted for many different specific tasks.

Diffusion Model A type of AI used for generating images by gradually refining random noise into coherent pictures.

Getting Help​

Don't see a term you're looking for?

  • Check the specific lesson where you encountered it for context
  • Search online for "[term] AI meaning" for additional explanations
  • Ask in AI communities - most people are happy to help explain concepts

Remember: AI terminology evolves quickly. New terms appear regularly as the technology advances. Focus on understanding core concepts rather than memorizing every definition.