Everyone on your team uses AI. Almost none of them are actually AI-literate.

You keep hearing that your people use AI, and you keep hiring graduates who say they know it. And yet the outputs are strangely uneven: one person's AI-assisted work is sharp and original, the next person's is confident nonsense, and you cannot always tell in advance which you are going to get. The reason is a distinction most organisations have not yet drawn. Using AI and being good at AI are different things, and you are probably screening and staffing on the first when the second is what matters.

Here is the shape of it. Access is no longer the issue. In McKinsey's 2025 workplace research, 94% of employees and 99% of leaders are already familiar with generative AI, so nearly everyone is "using" it in some form. What has not kept pace is understanding. A 2024 Bank of England and Financial Conduct Authority survey found that while 75% of firms use AI, only about a third claim to fully understand the technology they are running. That space between using and comprehending is where the real competency, AI literacy, lives, and it is emerging as the new definition of a work-ready professional.

What is AI literacy, as distinct from just using AI?

It is three capabilities, not one. First, the ability to direct AI: to frame a task well, to design the loop it works in rather than fire off a vague prompt. Second, the ability to judge its output: to spot the plausible-but-wrong answer, to know what the model is likely to get right and where it will quietly mislead. Third, the ability to use it responsibly: to understand privacy, bias and the limits of what should be delegated. Someone who can do all three is AI-literate. Someone who simply pastes a prompt and ships the result is not, however confidently they use the tool.

Recent scholarship makes exactly this case, framing GenAI literacy as the competency that now defines work-readiness, and noting that how professionals actually use generative AI varies enormously from sector to sector. The headline for a leader is uncomfortable but useful: "we all use AI" tells you almost nothing about whether your organisation is any good with it.

Why does this gap matter so much right now?

Because you are making real decisions on the wrong signal. You hire a graduate because they "know AI", when what you needed to test was whether they can judge an AI output under pressure. You assume your team is capable because they are all active users, when activity and skill have come apart. And in some markets the literacy deficit is structural: in Australia, for instance, AI use is widespread while formal training and understanding remain low, part of the same pattern that leaves the country ranking last in the world for AI trust. High usage sitting on thin literacy is not readiness; it is exposure.

It is the same trap as why most organisations fail at AI adoption: the tool is present and the capability to use it well is assumed rather than built. Access was the easy part, and it is finished. Literacy is the part that still separates one organisation from another.

How do I build AI literacy rather than assume it?

By treating it as a defined competency you develop on purpose, not a trait people either have or lack.

1. Separate "uses AI" from "is AI-literate" in hiring. Ask a candidate to critique an AI output, not just produce one. The ability to find the flaw is the signal you actually want.
2. Teach the three literacies explicitly. Directing AI, judging its output, and using it responsibly. Employees are asking for exactly this training; McKinsey found the demand for it is real and largely unmet.
3. Make good practice visible and shared. Capture how your most effective people actually use generative AI in their roles, across functions, and spread it, since the how varies widely by sector.
4. Close the comprehension gap, not just the access gap. People already have the tools. Give them the understanding of when to trust the output and when not to, which is where most of the risk and the value sit.
5. Fund it as work-readiness, not a perk. AI literacy is becoming a baseline competency, so resource it like one rather than leaving it to individual initiative.

What does this change for me as a leader this quarter?

It changes the question you ask about your workforce. Not "are we using AI", which you have answered many times over, but "can our people use it well, and how would we know". That reframes hiring, training and assessment around judgement rather than activity, and it quietly upgrades what a work-ready person means in your organisation.

The shift is the same one running through the end of business as usual: as the tools become universal, the advantage moves to the humans who use them with skill and judgement. Build that literacy deliberately, in the people you have and the people you hire, and you get a workforce that is genuinely ready, rather than merely equipped.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.