AI Literacy for Everyone

Distinguish AI, automation, machine learning and generative AI. Identify familiar workplace examples and explain why a rules-based process is not necessarily AI.

Explore training data, tokens, prediction and context windows in plain language. Understand that fluent responses are generated outputs, not guarantees of truth.

Identify suitable tasks and common limitations. Compare a useful draft with an answer that sounds confident but lacks evidence, and decide when a human should take over.

Understand conversational assistants, copilots, research tools, image generators, coding assistants and enterprise systems. Choose a tool category by task and approved access.

Build prompts using a goal, relevant context, constraints, examples and an output format. Practise asking for missing information instead of encouraging invented details.

Improve a weak prompt through follow-up questions and refinement. Use critique as a starting point for review, and create a reusable prompt structure.

Practise drafting, summarising, brainstorming, extracting information and preparing meetings. Choose low-risk tasks where checking the result is practical.

Recognise invented facts, sources, quotations and calculations. Verify important claims against original evidence and distinguish uncertainty from a supported conclusion.

Identify personal, customer and company information. Apply the organisation's approved-tool rules and practise replacing sensitive inputs with fictional examples.

Explore fairness, transparency, accountability and intellectual-property questions through workplace cases. Identify who must review an output before it affects another person.

Challenge assumptions, compare explanations and identify missing evidence. Treat AI as a thinking partner while retaining responsibility for the judgment.

Break a larger task into research, analysis, drafting, review and improvement. Define what a reliable output looks like at each stage.

Distinguish a single prompt, reusable workflow, automation and agent. Use a simple example to identify permissions, checkpoints and possible failures.

Describe your role and recurring tasks. Select five to ten relevant use cases, then rank them by value, risk, effort and ease of checking the result.

Choose a recurring work task and design an AI-assisted process with approved inputs, clear prompts, verification and a human approval step. Test it on a fictional or authorised example.