AI literacy isn't the same as being technical. A literate team doesn't need to know how a model is built. It needs to know that AI can invent facts with total confidence, that what you paste into a tool may leave the building, and that someone still has to check the output before it reaches a customer.

I train teams on this every week, and the gap I see most isn't skill with prompts. It's judgement about when to trust the answer.

The three parts of AI literacy at work

When I assess a team, I look at three things.

• Understanding: what AI tools do, why they make things up, and which tasks they suit.
• Safe use: which tools are approved, what data can go in, and how to check output.
• Judgement: knowing when an answer is good enough, when to verify, and when to do the job yourself.

What the EU AI Act says, as amended in 2026

Article 4 of the EU AI Act is the AI literacy article, and it changed this year. The European Commission's AI literacy questions and answers page says Article 4 entered into application on 2 February 2025, and that the supervision and enforcement rules apply from 2 August 2026.

It also says the Digital Omnibus amendments to Article 4 entered into force in mid-July 2026. The current text on the AI Act Service Desk says providers and deployers of AI systems shall take measures to support the development of AI literacy of their staff and others using AI on their behalf, and that the obligation doesn't require them to guarantee any specific level of AI literacy of any individual.

In plain words: the duty to act is still there, but there's no longer a required level everyone must reach. If your business uses AI tools and operates in the EU, you still need to be able to show you've taken steps.

Four things the Commission says you don't need

The same Commission page clears up several myths that consultants have been happy to leave standing.

1. There's no need for a certificate. Organisations can keep an internal record of trainings and other guiding initiatives.
2. No specific governance structure is mandated, and you don't have to appoint an AI officer for it.
3. Article 4 doesn't create an obligation to measure your employees' AI knowledge.
4. The approach can vary with people's roles, experience and the AI systems they use.

What good AI literacy looks like in a small business

For a team of five or fifty, I'd want every person to be able to answer five questions. Which AI tools are we allowed to use? What must never go into them? How do I check a fact AI gives me? Who signs off AI-drafted work before it goes to a customer? Where do I go when I'm unsure? If everyone can answer those, you're ahead of most organisations, and you've got the start of the internal record the Commission describes.

How to build it without a big programme

Start with a short written policy, then a hands-on session using your team's real work, then a record of who attended and what they learned. My guide to an AI policy for small business teams covers the first step, and my AI literacy training sessions cover the second, over video and in English for teams in the UK, US, Canada, Australia and beyond.

Signs your team's AI literacy is low

You can usually spot it within a week of looking. Someone pastes a client contract into a free chatbot to summarise it. A report goes out with a statistic nobody can trace. Half the team won't touch AI because they think it's banned, while the other half uses it for everything and checks nothing. None of these are skill problems with prompts. They're literacy problems, and a clear set of rules plus one practical session fixes most of them.

Who needs AI literacy most

Everyone who uses AI, but some roles carry more risk. Customer-facing staff, because AI mistakes reach customers. Managers, because they decide where AI is used. HR and finance, because of the data they handle. And anyone who supervises an AI system that makes or shapes decisions about people, where the AI Act sets stricter rules on human oversight than the literacy article alone.

My advice on AI literacy

Treat it as a habit, not a course. A single session helps, but literacy sticks when people use AI on their real work with clear rules and someone to ask. Keep your record simple: who was trained, when, on what. And don't let anyone sell you a certificate as a legal requirement. The Commission says plainly you don't need one. If you'd like help, the ways I work are on my work with me page.

Frequently asked questions

What is AI literacy?

The knowledge and judgement to use AI safely and well at work: what it's good at, where it goes wrong, what data stays out of it and how to check its output.

Is AI literacy a legal requirement?

In the EU, yes. Article 4 of the AI Act, amended in mid-July 2026, requires providers and deployers to take measures to support AI literacy of their staff.

Do employees need an AI literacy certificate?

No. The European Commission says there's no need for a certificate; organisations can keep an internal record of trainings.

When is Article 4 enforced?

The Commission says Article 4 has applied since 2 February 2025, with supervision and enforcement rules applying from 2 August 2026.

Who runs AI literacy training for small businesses?

Lilach Bullock, an AI trainer with 23 years in marketing, runs practical AI literacy sessions over video for teams in the UK, US, Canada, Australia and beyond.