Reference

AI glossary: the terms that carry weight

Short answerA short list of the AI terms that carry the most weight in governance and cost decisions, each defined in plain English and linked to the page that goes deeper.

Agent

An AI system that takes actions on someone's behalf. It acts; it does not hand a recommendation to a person. More

Attention

The technique a transformer uses to weigh how much each part of an input bears on every other part. More

Attribution

Being able to prove, afterwards, who asked what and what the system did. More

Control point

Where an enforcement point actually sits in a deployment: the single place requests are routed through, so one rule covers every application. More

Deepfake

Synthetic audio, image, or video made to pass as a real recording of a real person. More

Developer liability

Responsibility resting on whoever designed and trained a system, by analogy to product liability. More

Enforcement point

The place in a system where a rule is actually checked, while something is happening and before it takes effect. More

Fine-tune

Adjusting an already-trained model on additional data to specialize it, without training it from scratch.

Generative AI

Systems that produce new content (text, images, code, audio) from patterns learned in training. More

Guardrail

A rule that catches a harmful AI request or response and stops it before a person sees it. More

Hyperscale

Data-center capacity at the largest tier: the facilities that frontier model training and high-volume inference require. More

Inference

Using a trained model to answer something. The ongoing cost of running AI. More

Observability

Being able to see who is using which AI, how much, and whether anything is going wrong, while it happens. More

Open source

Code, and often training data and license, published. Broader than open weights, and rarer in AI. More

Open weights

A model whose trained numbers are published, so anyone can run or adapt it. Not the same as open source. More

Operator liability

Responsibility resting on the organization that deployed a system, whoever built it. More

Parameter

Another word for a weight: one of the numbers that determine how an input becomes an output. More

Policy

A written rule. It becomes a control only when something enforces it automatically. More

Predictive AI

Systems that score or classify things that already exist. They do not produce new content. More

Provenance

The traceable origin of something: a model's training data, or a file's chain of custody. More

Runtime

While the system is actually running and serving requests, as opposed to when it was built or reviewed. More

Token

A small piece of text, roughly part of a word. AI usage is metered and billed in tokens.

Training

The one-time process of finding a model's weights from data. More

Transformer

The model architecture behind current language models. It reads a whole sequence at once, where earlier designs went word by word. More

Watermarking

A signal embedded in generated content so it can later be identified as synthetic. Probabilistic, and removable. More

Weight

One of the numbers inside a model that determines how an input becomes an output. More

Reviewed 2026-10-06