Glossary — AI terms explained plainly, by Nicole Junkerman

Shared vocabulary makes the rest of the guide easier to read. These short definitions stay practical and broad, so a newcomer and a working team can use the same words to describe the same things.

Model: A system trained to find patterns and produce outputs from inputs. In everyday use a model might classify information, answer questions, summarise text, or suggest next steps. It does not understand context the way a person does, so its output still needs review.
Prompt: The instruction, question, or context given to an AI system. Better prompts describe the task, the audience, the constraints, and the desired format. Prompting helps, but good results also depend on data, system design, and whether the task suits AI at all.
Automation: A task performed with limited human action. Automation can save effort on repetitive work, but it should be matched to the right tasks and paired with checks where accuracy and fairness matter.
Oversight: The review and responsibility that keep people accountable for AI-assisted work. Effective oversight is proportionate: light for low-risk tasks, and stronger where decisions affect people, money, access, or rights.
Governance: The decisions, policies, review points, and responsibilities that shape how AI is used. Good governance answers ordinary questions: who owns the system, who checks outputs, what data is used, and what happens when something goes wrong.
Bias: A pattern in data or design that can make outputs unfair or skewed. Recognising bias is part of responsible use; it is one reason human review and clear sources remain important.

AI terms, explained without theatre.

Keep reading across the guide

AI terms, explained without theatre.

Keep reading across the guide

Stay informed