AI literacy is not a checkbox: learning to make informed decisions about AI

Most of us use AI in our everyday work. But using an AI tool is not the same as understanding it in a business context. How do you assess what AI can do, recognize its limitations and decide when its use is appropriate? This is the essence of AI literacy, a skill that is becoming increasingly important for professionals and organizations.

The relevance of AI extends far beyond technical roles. It supports professionals in writing and analyzing information, detecting patterns, assessing risks and making decisions. Sometimes its role is clearly visible, for example with a generative AI assistant. In other cases, AI operates behind the scenes as part of a larger system.

Jos van Hillegersberg, Academic Director of JADS and Professor of Business Information Systems, has studied the development and application of AI and digital technologies for several decades. For him, the broader use of AI does not mean that professionals should embrace every new application without question. Nor should they dismiss the technology because it is imperfect.

“Study the possibilities of AI, remain critical and follow its development. Don’t be too quick to declare something an absolute game changer, but don’t dismiss it too quickly either. Think in terms of possibilities, but remain critical about what it can and cannot do and where you can use it.”

That ability to explore, question and make informed choices is at the heart of AI literacy.

What does the EU AI Act require?

The EU AI Act has made AI literacy an explicit responsibility for organizations. Article 4 has applied since 2 February 2025 and was revised in July 2026.

Under the current provision, organizations that provide or deploy AI systems must take measures to support the development of AI literacy among employees and others who operate or use AI systems on their behalf. In doing so, they should consider people’s existing technical knowledge, experience, education and training, the context in which the system is used and the people who may be affected by it.

The AI Act does not prescribe one specific level of AI literacy that every individual must attain. This is an important distinction. A communications professional using generative AI to draft a text needs a different kind of understanding from a data scientist developing a predictive model or a manager responsible for AI-supported decisions.

AI literacy therefore cannot be reduced to one general training course for an entire organization. What people need to know depends on their role, the application and the potential consequences if something goes wrong. Training can be part of an organization’s approach, but AI literacy also needs to be reflected in its working practices, responsibilities and decision-making.

More than understanding the technology

Technical knowledge matters, but it is only part of the picture. Dr Danielle Sent is Associate Professor and Program Director Professional Education at JADS. Her research focuses on trustworthy AI, decision-support systems and the interaction between people and AI.

Sent describes AI as something that can fulfil different roles. It can support a professional as an assistant, act as a critical colleague that points out something they may have missed, or take over a clearly defined task. Determining which role is appropriate requires an understanding of both the technology and the situation in which it is used.

“It is never just about the model itself. It is about the model, the person using it and the environment in which it is used. You can have a very good model, but if the user handles it incorrectly or the organization does not have its procedures in order, you will run into problems.”

Responsible use therefore cannot be judged solely by asking whether an AI system produces accurate results. Organizations also need to consider how its output is interpreted, who checks it, what happens when circumstances change and who remains responsible for the eventual decision.

The stakes matter too. Using AI to improve an internal email is different from using it to assess a job applicant, detect fraud or support a medical decision. As Sent explains, there is no single answer that works for every application. The appropriate approach depends on who uses the system, in what setting, for what type of decision and whose interests are affected.

Four questions every professional should ask about any AI Model

You do not need to understand every technical detail of an AI model to use it responsibly. You do need to ask sensible questions about its purpose, limitations and consequences:

1. What am I using AI for?
   Is it supporting me, challenging my thinking or performing a task on my behalf?
2. What might the system miss or get wrong?
   Consider the quality of the available information, possible bias and relevant context that the system may not take into account.
3. What are the consequences if the output is wrong?
   The greater the potential impact on people, the more scrutiny and human oversight will be needed.
4. Who checks the result and remains responsible?
   An AI-generated recommendation does not remove professional or organizational responsibility.

These questions also help professionals use AI more actively. AI does not have to replace critical thinking. It can be used to test an assumption, uncover counterarguments or identify evidence that challenges your initial position. Used in that way, it can sharpen your thinking rather than simply do the thinking for you.

Why AI literacy is never finished

One of the most persistent misconceptions is that AI literacy can be addressed once: through a training course, a set of guidelines or an approved-tools list. In reality, all three can become outdated.

AI systems are regularly updated and may behave differently after an update. New tools enter the workplace, sometimes before the organization has formally assessed them. An application originally introduced for a low-impact task may gradually be used for more consequential decisions. Employees develop new routines and may begin to rely on automated output more heavily. Data, user groups and working environments change, which can affect the quality and relevance of results.

Legislation and official guidance are developing too. Article 4 itself was amended little more than a year after it first became applicable. Recommendations, practical examples and expectations around AI literacy will continue to evolve as organizations and regulators gain experience.

Uncertainty is therefore not a temporary problem that will disappear once AI systems become more advanced. It is a permanent part of working with a technology that continues to change, learns from changing data and is applied in new contexts. New capabilities create new possibilities, but can also introduce risks that were not considered when an organization first developed its policies or training.

Organizations must therefore keep asking whether employees still have the appropriate knowledge, whether procedures still fit the way AI is actually being used and whether existing safeguards remain effective. Monitoring matters not only because models can change, but also because people and organizations change around them.

Staying curious and staying in control

For Van Hillegersberg, the skills professionals need in response to AI are not entirely new. Curiosity, critical thinking, a willingness to keep learning and the ability to reflect on consequences have always been important within academic education. AI makes them even more relevant.

“What are the most important skills we try to teach people at a university? Be curious, look at new things, keep learning, be critical and keep thinking. Remain in control. Be enthusiastic about technology, but also consider what it means for people and organizations.”

This reflects the JADS approach to AI literacy. The aim is not simply to teach people how to operate today’s tools. Professionals need enough understanding to evaluate new developments, place them in context and decide whether an application genuinely adds value.

That requires knowledge of data and technology, but also attention to organizations, human behaviour, ethics and implementation. Above all, it requires an ongoing conversation about how AI is being used, where it adds value and where additional knowledge, safeguards or human judgement are needed.

For professionals, this starts with remaining curious without becoming uncritical. Experiment, but understand the purpose. Use AI, but continue to assess its output. And remain responsible for the choices you make with it.

Stay informed and keep asking the right questions

AI technologies, professional applications and European regulation continue to develop. Stay up to date with JADS insights that help you understand what these changes mean for your work and organization.

The JADS newsletter brings you academic perspectives on current developments in data and AI, new research and insights from JADS experts, practical cases from organizations, useful frameworks and tools, and invitations to relevant webinars, events and professional programs.