AI & Machine Learning terms

67 AI & Machine Learning terms explained in plain English.

Agent2Agent protocol (A2A)The Agent2Agent protocol (A2A) is an open standard that lets AI agents built by different vendors communicate with each other, share tasks, and exchange results.
Agentic WorkflowAn agentic workflow is a structured process in which an AI agent independently plans and executes a sequence of steps to complete a goal, using available tools and adapting its plan based on intermediate results.
AI BenchmarkingAI benchmarking is the process of evaluating AI models using standardised tests and datasets to measure their performance on specific capabilities -- reasoning, language understanding, mathematics, coding, factual knowledge, and safety.
AI Content DetectionAI content detection refers to tools and techniques designed to determine whether a piece of content -- text, image, video, or audio -- was generated by an artificial intelligence model rather than a human.
AI CopilotAn AI copilot (also called an AI assistant) is a software layer built on a large language model or similar AI system that works alongside a person to help them get things done faster.
AI EthicsAI ethics is the application of ethical principles -- fairness, transparency, accountability, privacy, and human dignity -- to the design, deployment, and governance of artificial intelligence systems.
AI guardrailsAI guardrails are rules, filters, and checks placed around an AI system to keep it safe, on-topic, and within policy.
AI JailbreakingAI jailbreaking refers to techniques that attempt to make an AI model produce outputs it would normally refuse -- such as generating harmful content, providing instructions for dangerous activities, or abandoning its assigned persona or restrictions.
AI Red-TeamingAI red-teaming is the deliberate, adversarial testing of AI systems by a team whose explicit goal is to find failures, harmful outputs, safety vulnerabilities, and capability limitations before the system is deployed.
AI WatermarkingAI watermarking is the practice of embedding hidden signals into AI-generated content -- text, images, audio, or video -- that allow the content to be identified as AI-generated when later inspected.
AlgorithmAn algorithm is simply a set of step-by-step instructions that tells a computer how to complete a task.
Anomaly DetectionAnomaly detection is the process of identifying data points, events, or observations that deviate significantly from an established baseline of normal behaviour.
Artificial IntelligenceArtificial intelligence, or AI, is when a computer or machine is designed to do things that normally require human thinking.
AutomationAutomation is when technology performs tasks with little or no human involvement.
Autonomous AgentsAutonomous agents are AI systems that can pursue a goal over a series of steps, making decisions, using tools, and adapting their approach based on intermediate results -- without requiring a human to approve every individual action.
Bias in AIBias in AI refers to situations where an artificial intelligence system produces unfair or skewed results because of problems in its training data, design, or assumptions.
Chain-of-ThoughtChain-of-thought (CoT) prompting is a technique for improving the reasoning quality of large language models by asking them to generate intermediate reasoning steps before producing a final answer, rather than jumping directly to the answer.
ChatbotA chatbot is a computer program that you can have a conversation with, usually by typing messages.
Computer-use agentA computer-use agent is an AI system that controls a computer or web browser the way a person does: it looks at screenshots, decides what to click or type, and repeats until a task is done.
Context windowA context window is the limit on how much text an AI language model can work with at one time, counting your instructions, any documents you attach, the conversation so far, and the model's own answer.
Data poisoningData poisoning is an attack where someone deliberately slips bad examples into the data used to train or fine-tune an AI model, so the model learns harmful behaviour, makes targeted mistakes, or responds to a hidden trigger.
Differential PrivacyDifferential privacy is a rigorous mathematical framework for privacy protection that works by adding carefully calibrated random noise to data, query results, or model outputs.
Diffusion modelA diffusion model is a type of generative AI that creates content by starting with random noise and gradually removing it, step by step, until a clear image, video, or sound appears.
EmbeddingsEmbeddings are lists of numbers, called vectors, that an AI model produces to represent the meaning of a piece of content such as a word, sentence, image, or document.
Expert SystemAn expert system is an AI program that encodes the knowledge and decision-making logic of a human expert in a specific domain as a set of 'if-then' rules, and applies those rules to make decisions, diagnoses, or recommendations.
Explainable AI (XAI)Explainable AI (XAI) refers to a set of methods, techniques, and design principles that make the outputs and decision-making processes of AI systems understandable to the humans who use, oversee, or are affected by them.
Federated LearningFederated learning is a machine learning approach in which a model is trained collaboratively across many devices or data silos without the raw training data ever leaving those devices.
Function CallingFunction calling (also called tool use or tool calling) is a capability of modern large language models that allows them to invoke external functions, APIs, or tools as part of generating a response.
Generative AIGenerative AI is a type of artificial intelligence that can create new content — including text, images, music, code, and video — based on instructions or examples you give it.
Hallucination (AI Hallucination)An AI hallucination is when an artificial intelligence tool generates information that sounds confident and believable but is actually wrong or completely made up.
Inference CostInference cost refers to the computing resources -- CPU, GPU, memory, network bandwidth, and energy -- consumed every time a trained AI model generates a prediction, response, or output.
Knowledge GraphA knowledge graph is a structured way of storing information that represents real-world entities (people, places, organisations, concepts) and the relationships between them as a connected network.
KV cacheThe KV cache, short for key-value cache, is memory an AI language model uses to store calculations about earlier tokens in a conversation so it doesn't have to repeat them for every new word it writes.
Large Language ModelA large language model, or LLM, is a type of artificial intelligence that has been trained on enormous amounts of text so it can understand and generate human language.
LLM evaluation (evals)LLM evaluation, often called evals, is the practice of testing an AI application with a set of realistic inputs and scoring its outputs for accuracy, safety, tone, and other qualities.
LLM-as-a-judgeLLM-as-a-judge is a technique where an AI language model grades the outputs of another AI system, using instructions or a rubric that describe what a good answer looks like.
LoRA (low-rank adaptation)LoRA, short for low-rank adaptation, is a technique for customizing an AI model by training a small set of extra weights instead of changing the whole model.
Machine LearningMachine learning is a way for computers to learn from experience, much like how you get better at something the more you practice.
Mixture of experts (MoE)A mixture of experts (MoE) is an AI model design that divides part of the network into many smaller 'experts' and uses a router to send each token to only a few of them.
Model collapseModel collapse is a decline in AI model quality that can happen when models are trained repeatedly on content made by other AI models instead of by people.
Model Context Protocol (MCP)The Model Context Protocol (MCP) is an open standard for connecting AI assistants to outside tools and data, such as files, calendars, databases, and business apps.
Model DistillationModel distillation (or knowledge distillation) is a technique for creating a smaller, more efficient AI model (the 'student') that is trained to approximate the behavior of a larger, more capable model (the 'teacher').
Model DriftModel drift is the phenomenon where a machine learning model's performance degrades over time because the real-world data it encounters in production differs from the data it was trained on.
Neural NetworkA neural network is a type of computer system inspired by how the human brain works.
On-Device AIOn-device AI refers to machine learning models that run locally on a device -- a smartphone, laptop, smart speaker, or embedded sensor -- rather than sending data to a cloud server for processing.
Open-weight modelAn open-weight model is an AI model whose trained parameters, or weights, are released publicly so anyone can download, run, and often fine-tune it on their own hardware.
Prompt (AI Prompt)A prompt is the instruction or question you type into an AI tool to tell it what you want.
Prompt cachingPrompt caching is a feature of AI APIs that saves the processed version of the beginning of a prompt, such as long instructions, a big document, or tool definitions, so later requests that start the same way don't have to process it again.
Prompt InjectionPrompt injection is a class of attack against AI language models in which malicious instructions are embedded in data that the AI is directed to process -- such as a document, email, web page, or user message -- causing the AI to override its original instructions and follow the attacker's instructions instead.
QuantizationQuantization is a model compression technique that reduces the size and memory requirements of a neural network by representing its weights and activations in lower-precision numerical formats.
Reasoning modelA reasoning model is a large language model trained to 'think' before it answers, generating an internal chain of reasoning steps, checking its work, and trying different approaches before giving a final response.
RerankingReranking is a second pass in a search system: a fast first step finds a broad set of possibly relevant results, then a more careful AI model reads each one alongside the query and reorders them by how well they actually answer it.
Responsible AIResponsible AI is a broad term for the organizational practices, governance frameworks, technical standards, and cultural commitments that ensure AI systems are built and used in ways that are safe, beneficial, fair, and trustworthy.
Scaling lawsScaling laws are patterns researchers found showing that AI language models improve in a predictable way as they're trained with more computing power, more data, and more parameters.
Small language model (SLM)A small language model (SLM) is an AI language model with relatively few parameters, typically from under a billion to a few billion or so, designed to run quickly and cheaply, sometimes directly on a phone, laptop, or small server.
Speculative decodingSpeculative decoding is a technique for making large AI models generate text faster.
Structured outputsStructured outputs is a feature that makes an AI language model produce responses in a precise, machine-readable format, usually JSON that matches a schema you define, with exact field names and data types.
Synthetic DataSynthetic data is data that has been artificially generated by a computer model rather than collected from real-world events or people.
Temperature (LLM sampling)Temperature is a setting that controls how much randomness an AI language model uses when picking each next word.
Text-to-ImageText-to-image is a category of generative AI that creates images from natural language descriptions (called prompts).
Text-to-VideoText-to-video is a generative AI capability that creates video clips from natural language descriptions.
Token PricingToken pricing is the billing model used by AI API providers like OpenAI, Anthropic, and Google to charge for large language model usage.
TokenizerA tokenizer is the part of an AI language model system that breaks text into tokens, small chunks such as whole words, parts of words, or punctuation, and converts them into numbers the model can process.
Training DataTraining data is the collection of examples that an AI system learns from, much like textbooks and practice problems for a student.
Vector databaseA vector database is a database built to store embeddings, the number lists AI models use to represent meaning, and to quickly find the ones closest to a query.
Vibe codingVibe coding is a way of building software where a person describes what they want to an AI tool in everyday language, accepts the code it writes, and keeps asking for changes until the result works, often without reading the code closely.
Voice CloningVoice cloning uses AI to analyze recordings of a real person's voice and produce a model that can generate new speech in that voice from any typed text.