Learn the terms used in artificial intelligence, machine learning, generative AI, language models, prompt engineering, agents, automation, research, safety, and responsible AI use.
Artificial intelligence can feel complicated because it introduces unfamiliar words, abbreviations, technologies, and systems.
This glossary explains common AI terms using clear language and practical examples. It is designed for beginners, authors, entrepreneurs, students, educators, creators, professionals, and everyday users.
Some terms describe the technology itself. Others describe how people communicate with AI, how systems use tools, how models are trained, or how users can identify risks and verify information.
A step-by-step mathematical or logical process used by a computer to solve a problem, analyze data, or complete a task.
A theoretical form of AI capable of learning and performing across a broad range of intellectual tasks at or near human ability.
The broad field of creating computer systems that perform tasks associated with intelligence, including language processing, prediction, recognition, planning, generation, and decision support.
AI designed to perform one task or a limited group of related tasks. Most AI systems currently in use are forms of narrow AI.
A machine-learning structure inspired loosely by networks of biological neurons. It contains connected layers that process information and learn patterns.
A hypothetical form of intelligence that would exceed human intellectual abilities across many areas.
An AI system designed to work toward a goal by planning steps, using tools, taking actions, checking progress, and adjusting when needed.
A conversational or task-oriented AI system that helps users answer questions, create content, analyze information, organize work, or use connected tools.
The use of AI to complete repeatable tasks, move information between systems, generate outputs, or support workflows with limited manual effort.
Unequal, distorted, or unfair patterns in AI outputs caused by training data, system design, evaluation methods, human assumptions, or missing perspectives.
An AI assistant embedded into a person's workflow to provide suggestions, drafts, analysis, code, summaries, or task support while the human remains in control.
The rules, standards, oversight, accountability, testing, permissions, and decision processes used to manage artificial intelligence responsibly.
A response in which an AI system generates information that sounds believable but is false, invented, unsupported, or unrelated to the available evidence.
The ability to understand how AI works, use AI tools effectively, recognize limitations, verify information, protect privacy, and make responsible decisions.
A trained mathematical system that processes input and produces predictions, classifications, recommendations, actions, or generated content.
The field of reducing harmful, unintended, unreliable, insecure, or uncontrolled behavior in artificial intelligence systems.
An application programming interface allows software systems to exchange information and use each other's features through defined commands.
The use of technology to complete tasks or processes with reduced manual effort. Automation may or may not include artificial intelligence.
A standardized test or collection of tasks used to compare the performance of AI systems or models.
Extremely large or complex collections of information that require advanced tools and computing methods to store, process, and analyze.
A software program designed to perform automated actions. A chatbot is a bot focused on conversation.
A software system designed to communicate with users through written or spoken conversation.
The process of assigning information to a category, such as identifying whether an email is spam or whether an image contains a cat or dog.
Artificial intelligence that analyzes and interprets images, video, scans, documents, objects, faces, or physical environments.
A numerical estimate indicating how certain a model is about a prediction or classification. A high score does not automatically guarantee correctness.
The information available to an AI system while generating a response, including instructions, conversation history, files, retrieved sources, and tool results.
The amount of text and other information a language model can consider during one interaction or conversation.
Related: Large Language ModelsLegal protection for original creative works. AI-generated and AI-assisted content may raise questions about authorship, training data, originality, and commercial use.
Information used by computer systems. Data may include text, images, audio, video, numbers, transactions, measurements, or human feedback.
An organized collection of information used to train, test, evaluate, or operate an AI model.
A form of machine learning that uses layered artificial neural networks to identify complex patterns in large amounts of data.
Related: AI vs. Machine LearningSynthetic or altered media that realistically imitates a person's face, voice, appearance, or actions.
A type of generative model commonly used to create images by gradually transforming noise into a structured visual output.
A numerical representation of information that allows AI systems to compare meaning, similarity, relationships, or context.
A skill or behavior that appears as an AI model becomes larger or more capable, even when that exact ability was not directly programmed.
A rule-based AI program designed to imitate decisions within a specialized field using structured knowledge and predefined logic.
The process of confirming whether a claim, date, quotation, statistic, source, or explanation is accurate.
Related: How to Verify AI InformationAdditional training that adapts a general model to a specific task, industry, format, style, or collection of examples.
A large general-purpose model trained on broad information that can be adapted to many tasks and applications.
Artificial intelligence designed to create new content such as text, images, music, video, voices, computer code, designs, or plans.
Related: What Is Generative AI?A model trained to produce new outputs that resemble patterns found in its training material.
Connecting an AI response to reliable sources, supplied documents, databases, search results, or verified information.
Rules, restrictions, permissions, filters, approval steps, or technical controls designed to reduce unsafe or unwanted AI behavior.
A false, unsupported, or invented AI response presented in a way that may sound confident or believable.
A workflow in which a person reviews, approves, corrects, or controls important AI-generated decisions or actions.
A process in which people and AI systems contribute different strengths to complete a project or solve a problem.
The process of using a trained AI model to analyze new input and produce a prediction, answer, classification, recommendation, or generated output.
Information given to an AI system, such as a prompt, file, image, question, audio clip, command, or dataset.
Legal rights connected to creative works, inventions, brands, designs, trademarks, and other original creations.
An AI model trained on large amounts of language-related data so it can process prompts and generate text, summaries, explanations, translations, code, and conversations.
Training examples that include the correct answer, category, or description.
The amount of time between sending an AI request and receiving a response or completed action.
A branch of artificial intelligence in which systems identify patterns from data and improve performance through examples.
Information an AI system retains or retrieves to support future interactions, preferences, instructions, progress, or task continuity.
A trained mathematical system that transforms input into an output.
The process of exposing a model to examples, measuring errors, and adjusting internal mathematical relationships so performance improves.
Artificial intelligence that can process or generate more than one kind of information, such as text, images, audio, video, and files.
A process that combines several AI tools or models, with each tool handling the part of a project it performs best.
Related: Multi-AI WorkflowsThe field of AI focused on helping computers process, understand, analyze, and generate human language.
A layered machine-learning structure made of connected mathematical units that learn patterns from data.
Tools and platforms that allow people to build applications, websites, automations, or digital systems without traditional programming.
An adjustable mathematical value within an AI model that helps represent patterns learned during training.
An estimate produced by an AI or machine-learning model about a likely category, result, value, event, or next token.
Artificial intelligence used to estimate future outcomes, detect patterns, classify information, or calculate probabilities.
The question, instruction, description, example, file, or command given to an AI system.
The practice of designing clear instructions, context, examples, formats, requirements, and limits to improve an AI response.
Related: Prompt EngineeringAn attempt to manipulate an AI system by inserting instructions that override, weaken, or conflict with its intended rules.
A reusable prompt structure containing placeholders that can be updated for different topics, audiences, goals, or projects.
Related: How to Write Better PromptsA method that retrieves relevant information from outside sources and provides it to a generative model before the response is created.
The process of analyzing relationships, following steps, comparing evidence, solving problems, or selecting among possible actions.
An AI system that ranks or suggests products, videos, music, books, services, or information based on patterns and preferences.
A learning method in which a system improves through actions, feedback, rewards, and penalties.
The design and use of artificial intelligence with attention to accuracy, fairness, transparency, privacy, security, accountability, and human control.
The field of designing and controlling machines that perform physical tasks. Robotics may combine sensors, software, AI, planning, and movement.
Technology that converts spoken language into text or structured commands.
A machine-learning method in which a model learns from examples containing correct labels or expected answers.
Artificially generated information designed to imitate the structure or patterns of real-world data.
Hidden or platform-level instructions that guide an AI system's behavior, role, priorities, tools, and response rules.
A setting that influences how predictable or varied a generative model's output may be. Higher settings generally allow more variation.
A system that creates images from written descriptions or prompts.
Technology that converts written text into spoken audio.
A small unit of language processed by an AI model. A token may be a complete word, part of a word, punctuation mark, number, or symbol.
Related: Large Language ModelsInformation used to teach an AI model patterns, relationships, structures, or desired outputs.
A neural-network architecture designed to process relationships within sequences of information. Transformers are widely used in large language models.
The ability of an AI system to use search, files, code, databases, email, calendars, websites, or other connected applications.
The process of adjusting a model through examples and feedback so it becomes better at performing a task.
A machine-learning method in which a model searches for patterns, groups, or relationships without being given correct labels.
The instruction or question provided directly by the person using an AI system.
The process of checking whether an AI model or system performs effectively on information that was not used to train it.
A database designed to store and search numerical representations of meaning, similarity, and relationships.
The process of checking an AI-generated claim or output against reliable evidence, authoritative sources, testing, or expert review.
Technology that creates synthetic speech designed to imitate the sound and speaking characteristics of a particular voice.
Artificial intelligence, algorithms, models, data, inference, prediction, and automation.
Explore AI foundationsTraining data, supervised learning, unsupervised learning, reinforcement learning, neural networks, and deep learning.
Explore machine learningGenerative models, text-to-image, diffusion models, multimodal AI, synthetic content, and content creation.
Explore generative AITokens, context windows, transformers, prompts, embeddings, system instructions, and language generation.
Explore language modelsPrompts, templates, context, examples, requirements, constraints, follow-up prompts, and prompt injection.
Explore prompt engineeringGoals, planning, memory, tool use, actions, permissions, autonomy, guardrails, and human approval.
Explore AI agentsHallucinations, bias, privacy, deepfakes, transparency, governance, verification, and accountability.
Explore responsible AIAPIs, automation, AI assistants, copilots, multi-AI systems, no-code platforms, and connected tools.
Explore AI toolsArtificial intelligence is the broader field. Machine learning is one method used within AI.
A chatbot usually answers. An agent may plan, use tools, take actions, check progress, and continue.
Predictive AI estimates an outcome. Generative AI produces new content.
Knowing the vocabulary makes it easier to compare tools, write better prompts, understand AI limitations, and follow new developments.
The AI Knowledge Center connects each major term to practical guides, examples, learning paths, books, tools, and responsible-use principles.
AI means artificial intelligence, the broad field of creating computer systems that perform tasks associated with human intelligence.
LLM means large language model, an AI model trained to process and generate human language.
A model is the trained mathematical system. An AI tool is the product, application, or interface through which people use one or more models.
An AI hallucination is false or invented information generated in a way that may sound believable.
Artificial intelligence develops quickly, and companies, researchers, governments, and users may apply terms differently as new systems and capabilities appear.
Explore common questions about AI tools, prompts, accuracy, privacy, work, publishing, agents, safety, and the future.