AI training:
giving human capabilities
their rightful place

Setting a goal, interpreting a situation, questioning a proposal, making a judgement call: these capabilities determine the quality of work done with AI. Developing them deserves an explicit place in training programmes.

Continuing education providers, higher education institutions, learning and development (L&D) and HR leaders in companies: we share a responsibility. Training people to use AI also means developing the human capabilities needed to steer it.

Human skills are on the rise

Coursera’s Global Skills Report 2026, drawing on data from Coursera and Udemy, ranks critical thinking and judgement at the top of the human skills it identifies. Enrolments related to critical thinking grew by 107% in one year. Ethics, governance and complex problem-solving also feature prominently. These figures describe learning behaviour on the platforms: interest in human capabilities is growing alongside interest in AI.

Source 01  /  Coursera · Global Skills Report 2026 (opens in a new tab)

This trend echoes the conviction behind Infinitum: the more tasks we hand over to AI, the more we need to work on the quality of our steering.

Three ways of working with AI

Not all professional activities lend themselves to the same way of using AI.

Structured processes can be broken down into ordered, repeatable steps: gathering information, sorting it, producing a document, sending it on. These situations lend themselves to automation, with defined rules and checks.

Hybrid processes also follow a sequence, but some steps call for judgement. Preparing a report, for example, may combine automated gathering and summarising with interpreting the results, weighing options and signing off. The challenge is to identify the moments when human intervention is needed.

Open questions require the path to be built as the thinking unfolds: setting a strategic direction, understanding a complex situation or imagining answers to a new problem. A dialogue with a chatbot can support this exploration. The user asks questions, tests the options against each other and adjusts their approach.

Human steering therefore takes different forms: setting the rules, stepping in at decisive moments or guiding open-ended thinking. Training must prepare people to recognise these situations and to draw on the right capabilities.

Capabilities in demand, but learning that remains too implicit

People are expected to be critical, creative and adaptable. But where do they explicitly learn to develop these capabilities?

In higher education, they can be part of a “hidden curriculum”: opportunities to exercise them exist, but students do not always know which capabilities they are working on or how to progress. This is the point made by Oscar Ybarra in an article published by AACSB in 2023, which argues for teaching adaptability intentionally.

Source 02  /  Oscar Ybarra · AACSB, 2023 (opens in a new tab)

In Houston, two public schools have chosen to give these capabilities a visible place in the timetable. In 2026, they are trialling a schedule that devotes mornings to academic learning and afternoons to projects that develop critical thinking, collaboration and adaptability, among other skills. Ulrich Boser describes the initiative in How Two Schools Are Leaning Into Human Skills in the Age of AI. He highlights one essential condition: linking these projects to the knowledge students acquire, because reasoning is always exercised on content and in context.

Source 03  /  Ulrich Boser · Forbes, 2026 (opens in a new tab)

This experiment illustrates an approach we advocate: making the development of human capabilities an explicit learning objective, with time, practice and feedback to make progress.

From intention to practice

“Developing critical thinking” becomes concrete when a participant learns to spot the assumptions behind an AI answer, test them against their own situation and decide what needs to be checked.

They still need to reuse this approach in their work. The research review Making soft skills “stick”, which gave rise to the COMPASS model, examines the conditions for this transfer. It highlights the combined role of capabilities, motivation, the work environment and the design of the training.

Source 04  /  Hamzah et al. · Published online, 2024 (opens in a new tab)

For institutions and companies alike, making this learning explicit therefore means planning for:

• capabilities identified in the learning objectives; exercises linked to real work situations; observation criteria and feedback; opportunities to reuse what has been learned.

This approach can inform AI training, but also job-specific learning paths, management and decision-making.

AI-IQ: a map to guide your thinking

Developed by Infinitum, the AI-IQ method brings together 28 human capabilities and 15 mindsets, along with a navigation map for thinking.

Why combine capabilities and mindsets? Because knowing how to do something does not guarantee that you will do it. Someone may know how to check information and still accept an answer without examining it. They may know how to explore several options and still stop at the first one.

With AI-IQ, users make their need and intention explicit, then choose the capabilities and mindsets that suit their situation. They can explore, compare, reframe, go back or check something more thoroughly. The map makes these choices visible, so they can be worked on.

This approach – Soft Skills Engineering for AI Usability – offers a framework for building this development into training programmes and employee learning paths.

Making room for human capabilities in programmes

Training institutions, L&D and HR leaders: let us give Human Skills in the Age of AI an explicit place.

Critical thinking, discernment, creativity, adaptability: these capabilities need to be translated into learning objectives, practical exercises and progression criteria.

Our ambition is for advances in artificial intelligence to go hand in hand with the development of human intelligence.