Vocational education and training has one particularity that sets it apart from other stages of education: its students enter the labour market within months, not years. What they learn today, they may be applying almost immediately in a workplace where AI is already present. That is why, when we talk about critical thinking and AI in VET, we are not talking about an abstract skill for the future, but about a professional survival skill for the very near present.
The problem is that many vocational programmes are still teaching students to carry out tasks that AI can already perform, while failing to teach them the one thing AI cannot do for them: make sound decisions about when to use it, how to check what it produces, and when to question its output. Training people to execute tasks in a world where execution is increasingly automated means preparing them to compete precisely where they are most likely to lose.
What does critical thinking mean when there is AI in the middle?
Critical thinking is not new. What has changed is the environment in which it is exercised.
In the past, critical thinking largely involved evaluating information that was relatively difficult to find and that came with more or less visible indicators of reliability. Now, information arrives instantly and with the appearance of authority, generated by systems that do not distinguish between what is true and what is merely plausible. Critical thinking today means having the discipline to question something that sounds flawless. And that discipline is harder to maintain, not easier, precisely because the answer appears complete.
In a VET classroom, this translates into something very concrete. An administration student asks AI to draft a contract; a computing student asks it to generate a piece of code; a healthcare student asks it about a protocol. All of them can receive an answer that looks correct. The skill we need to teach them is not how to generate that response — the machine can already do that — but how to determine whether they can trust it before taking it to a client, a system or a patient.
Why execution is no longer the differentiator
For decades, much of vocational education has been built around procedural competence: knowing how to process a payroll, install a system or apply a technique. This practical know-how still matters, but it is no longer what necessarily distinguishes a good professional from an average one, because an increasing number of these procedures can now be carried out with the support of AI tools.
What matters increasingly is the judgement required to oversee that execution: recognising when a payroll has been calculated incorrectly, when a system contains an error, when generated code introduces a security vulnerability, or when an automated text says something that does not apply to the situation. The valuable professional is not the one who executes fastest, but the one who knows when automated execution has gone wrong.
How does this translate into the classroom?
Teaching this does not require a new subject or specialised equipment. It requires a change in how tasks are designed.
Instead of asking students to produce a result, ask them to evaluate one generated by AI: here is a contract, a piece of code or a report produced by a tool — what is wrong with it? This exercise reverses the student’s role, moving them from executor to supervisor. And that is precisely the role they are likely to occupy in the workplace.
Another approach is to require students to document their verification process. Not just what the AI produced, but how they established whether it was correct, what they checked it against and what questions arose during the process. This is where critical thinking stops being an abstract concept and becomes an assessable habit.
This approach aligns with the international frameworks that are already shaping AI literacy. The OECD and European Commission’s AILit Framework, published in 2026, places particular emphasis on the ability to understand, manage and question AI, rather than simply create with it, which is still where much of the focus in education lies. VET has an advantage here that should not be overlooked: its practical orientation makes it a natural setting for teaching these principles through real workplace scenarios.
What is at stake for VET students?
A VET student who learns only to carry out automatable tasks enters the labour market competing with the very technology that automates those tasks. A student who learns to monitor, verify and make decisions occupies a very different position: they become the person overseeing the tool rather than simply performing the task that the tool can replace.
The difference between the two is not the technology they know, but the judgement they have developed to use it. And in VET, that judgement is not an additional skill sitting alongside the curriculum. It is increasingly what will determine how professionals work with AI over the next five years.
Frequently asked questions
How can I develop critical thinking about AI in a VET programme?
By changing the type of task. Instead of asking students to produce a result, ask them to evaluate one generated by AI and identify its flaws. Then require them to document how they verified the response, not simply what the response was. This shifts the student from executor to supervisor — a role that increasingly reflects how AI will be used in the workplace.
Isn’t it more important for students to learn how to use AI tools?
Learning how to use the tools is the easy part, and the tools themselves change every few months. What endures is the judgement required to decide when to trust their output and when not to. A professional who only knows how to operate the tool competes with it; a professional who knows how to oversee it remains responsible for the outcome.
What framework can I use as a reference?
The OECD and European Commission’s AILit Framework (2026) sets out AI literacy competencies and places particular emphasis on understanding, managing and questioning the technology, beyond simply creating with it. It provides a useful roadmap for vocational education and training.
Sources
- OECD and European Commission (2026). AILit Framework. Framework of literacy in artificial intelligence for education. Open access.
