Artificial Intelligence Glossary

Also: agentic AI · autonomous agent · tool use

An AI agent is a system that uses a language model to plan and execute multi-step tasks by calling external tools — searching, running code, querying APIs — and iterating on the results, rather than producing a single response.

Also: alignment · AI safety

AI alignment is the field concerned with ensuring AI systems reliably pursue the goals their operators actually intend, rather than the literal specification given or some proxy the system learned during training.

Also: evals · benchmarking · LLM evaluation

AI evaluation is the practice of systematically measuring model performance against defined tasks and criteria, using benchmarks, human review, or automated graders to quantify capability and detect regressions.

Also: AI regulation · AI policy · AI Act

AI governance refers to the policies, regulations, and internal controls that determine how AI systems may be developed and deployed, including risk assessment, transparency obligations, and accountability for harms.

Also: model red teaming · adversarial testing

AI red teaming is the practice of deliberately probing a model or AI-powered application for harmful, unsafe, or policy-violating behaviour before and after deployment.

Also: context length · context size

The context window is the maximum number of tokens a language model can consider at once, encompassing both the input prompt and the generated output.

Also: on-device AI · local inference

Edge AI is the practice of running AI models directly on end-user devices such as phones and laptops rather than on remote servers, keeping input data local to the device.

Also: vector embedding · text embedding · semantic vector

An embedding is a numerical vector representation of text, images, or other data, positioned in a high-dimensional space such that semantically similar items sit close together.

Also: finetuning · LoRA · PEFT · supervised fine-tuning

Fine-tuning is the process of further training a pre-trained model on a smaller, task-specific dataset to adapt its behaviour, style, or domain knowledge without training from scratch.

Also: grounded generation · citation grounding

Grounding is the practice of tying a model's output to specific verifiable source material, so that claims can be traced to documents rather than resting on the model's parametric memory.

Also: confabulation · AI hallucination

A hallucination is output from an AI model that is fluent and confident but factually incorrect or entirely fabricated, such as an invented citation, a non-existent API method, or a plausible-sounding but false statistic.

Also: model inference · serving · generation

Inference is the process of running a trained model to produce output for a given input — as distinct from training, which is the process of creating the model in the first place.

Also: AI jailbreak · safety bypass

A jailbreak is a prompt crafted to bypass an AI model's safety training and elicit output the model was trained to refuse, typically through roleplay framing, hypothetical scenarios, or obfuscated instructions.

Also: model distillation · teacher-student

Knowledge distillation is a training technique in which a smaller student model learns to reproduce the behaviour of a larger teacher model, yielding a compact model that retains much of the teacher's capability.

Also: LLM · language model · foundation model

A large language model (LLM) is a neural network trained on vast quantities of text to predict likely continuations of a sequence, producing systems capable of generating coherent language, answering questions, and writing code.

Also: MCP

The Model Context Protocol (MCP) is an open standard that defines how AI applications connect to external data sources and tools, providing a common interface so that any compliant client can use any compliant server.

Also: open weights · open source model · local model

An open-weight model is one whose trained parameters are publicly downloadable, allowing anyone to run, inspect, fine-tune, and deploy it on their own infrastructure without calling an external API.

Also: prompting · chain of thought · few-shot prompting

Prompt engineering is the practice of structuring input to a language model — through instructions, examples, formatting, and reasoning scaffolds — to reliably obtain the desired output.

Also: indirect prompt injection · LLM injection

Prompt injection is an attack in which malicious instructions are embedded in content an AI system processes — a web page, document, or email — causing the model to follow the attacker's instructions instead of the application's.

Also: model quantization · INT8 · 4-bit quantization

Quantization reduces the numerical precision of a model's weights — for example from 16-bit floating point to 8-bit or 4-bit integers — shrinking memory requirements and speeding inference at some cost to output quality.

Also: RAG · retrieval augmented generation

Retrieval-augmented generation (RAG) is a technique in which relevant documents are retrieved from an external knowledge source and supplied to a language model as context, so that its response is grounded in specific retrieved material rather than parametric memory alone.

Also: Reinforcement Learning from Human Feedback · preference tuning

RLHF (Reinforcement Learning from Human Feedback) is a training technique that aligns a model's behaviour with human preferences by training a reward model on human comparisons of outputs, then optimising the language model against that reward signal.

Also: vector search · neural search · hybrid search

Semantic search retrieves results based on meaning rather than keyword matching, using embeddings to find documents conceptually related to a query even when they share no literal terms.

Also: tokenization · tokenizer · BPE

A token is the atomic unit of text that a language model processes — typically a common word, a word fragment, or a punctuation mark — produced by splitting input text through a tokenizer before the model sees it.

Also: data provenance · dataset transparency

Training data provenance is documentation of where the data used to train a model came from, including its sources, licensing status, and any known rights disputes.

Also: transformer architecture · attention mechanism

The transformer is a neural network architecture, introduced in 2017, that uses self-attention to weigh the relevance of every input token against every other, and which underpins essentially all modern large language models.

Also: vector store · vector index · ANN search

A vector database is a data store optimised for indexing and querying high-dimensional embedding vectors, typically using approximate nearest neighbour algorithms to find the most semantically similar items quickly.