The knowledge and skills people need to use AI tools well and safely at work, including what the tools can and cannot do, how to check output, and what data may be entered.
Definition
AI literacy is the set of knowledge and skills that lets a person use, supervise, and question AI tools at work. It covers what a tool does and where it is likely to be wrong, how to check output against a source, what data may and may not be entered, how to spot hidden instructions in content, how to report a problem, and who to ask. Needs differ by role: everyone needs the basics; reviewers, use owners, and prompt and data stewards need more.
History
Calls for AI literacy grew with the spread of generative AI. Some laws now expect it; for example, the EU AI Act includes a duty on providers and deployers of AI systems to ensure a sufficient level of AI literacy among their staff. Check the current status and timelines of any rule that applies to you.
When to Use
Build AI literacy before a tool is rolled out, and refresh it when tools, rules, or roles change. Test it, as with any training, by looking at what people do, not only at attendance.
Step-by-Step
- List the audiences and the tasks each does with AI.
- Set the basics for everyone: limits of the tools, data rules, checking, injection awareness, and reporting.
- Add role-specific skills for reviewers, owners, stewards, and champions.
- Teach with real examples, including your own near misses.
- Practice checking, for example with seeded errors.
- Test and reinforce: measure behavior and results, not only completion.
Examples
- A short course for all staff on what may go into which tool.
- A reviewer workshop using seeded errors to practice spotting wrong output.
- A toolbox talk on hidden instructions in emails and documents.
Common Pitfalls
- One-time training that fades.
- Only teaching how to use the tool, not how to check it.
- Counting attendance instead of behavior.
- The same content for every role.