However, we live in a post-digital reality where technology is inextricably intertwined with how we think and learn. Technology actively shapes how we build knowledge. A one-sided focus on the tool and its output is therefore too limited. The real key lies not in the technology, but in the intention with which the learner uses it. That intention guides the conscious consideration we must make for each task and context: which thought processes do we outsource to AI, and which do we deliberately keep in our own hands? Making that choice is a skill in itself. The following four insights attempt to clarify the decisive role of that intention.

Insight 1: Self-Direction versus Self-Regulation

To grasp the role of intention, we must make a clear distinction between Self-Directed and Self-Regulated learning.

With Self-Directed Learning (SDL), you take charge of the broad strokes: you yourself decide what you learn and why. You set your own overarching learning goals.

With Self-Regulated Learning (SRL), you take control of the learning process, but the goal may be set by someone else (such as an assignment). You plan the approach, choose strategies, monitor your progress, and adjust where necessary.

One can be seen as a part of the other. Self-regulation (SRL) is a skill necessary for successful self-direction (SDL). After all, to effectively map out your own learning path (SDL), you must also be able to manage the process of getting there (SRL).

Insight 2: Learning Needs an Engine and a Compass

Every learning process needs an engine and a compass: motivation and metacognition.

Self-Determination Theory describes the engine. It states that every person has three basic psychological needs: autonomy (freedom of choice), competence (the feeling of being capable), and relatedness. When these needs are met, autonomous motivation arises: the will to learn from personal interest or conviction, not necessarily due to external pressure.

Metacognition is the compass. It is ‘thinking about your own thinking’: planning, monitoring, and adjusting your learning process. This metacognitive skill is essential to effectively use the motivational fuel and achieve your learning goals. Motivation and metacognition continuously reinforce each other.

Insight 3: Using AI is a Matter of Intention

The way we use AI is not an on-or-off switch. Its effectiveness depends entirely on the context and the goal. No single approach is inherently ‘good’ or ‘bad’; it’s about different intentions you can consciously adopt, such as:

  • The intention of convenience: The goal is to complete a task with minimal effort, driven by deadlines or the desire to move on quickly. This is cognitive offloading: the deliberate outsourcing of mental work. This becomes detrimental when the task itself was the learning objective.
  • The intention of exploration: Here, the goal is to experiment and discover. You explore a new topic, but also the possibilities and limitations of the AI tool itself. It’s a process of ‘tinkering,’ combining, and trying things out without a clearly defined end result.
  • The intention of deepening: The goal here is to enrich and challenge your own thought process. You use AI as a critical sparring partner to test your understanding, sharpen your arguments, or broaden your perspective, starting from your own prior knowledge.

To clarify with an example: consider using a tool like NotebookLM to summarize source texts. When you do this with texts you haven’t read yourself, your intention is pure efficiency. The thinking is outsourced, which, depending on your goal, detracts from your learning process. However, if you use the same tool on sources you do know, or even on your own notes, the intention shifts. It can then be a way to gain inspiration, see new connections, or support your own memory. The purely technical action is the same; the intention is different.

The discussion about ‘improper use’ of AI is therefore not about the tool, but about the (un)conscious choice of intention. This improper use, for example, is applying the efficiency intention where the deepening intention was required. The art is not to ban the tool, but to guide learners in this process. Because choosing the right intention is not a static decision. It is a dynamic and iterative process that requires self-regulation. For one and the same task, you plan your approach (which intention do I start with?), monitor the result (is this really helping me?), and adjust where necessary (perhaps for this sub-task, I need to switch from ‘convenience’ to ‘deepening’?). This continuous interaction, driven by metacognitive skills, forms the core of smart AI use.

Insight 4: Critical Thinking is Not Only a Prerequisite but Also a Consequence

A persistent misunderstanding is that you must first teach critical thinking as a separate skill before letting students work with AI. The flawed assumption is that they will then automatically be critical of the output. However, the logic is too simplistic: you cannot learn ‘critical thinking’ without situated knowledge, and the need to be critical also arises from the right learning intention.

A student with the intention of unbridled efficiency feels little drive to check the output of AI. After all, “good enough” is sufficient to check off the task. A student with the intention of deepening must critically evaluate the output, because uncritically adopting information would sabotage their own learning goal—deep understanding.

For the latter student, critical evaluation is not a separate skill but a logical consequence of self-regulation and metacognition in action. We need to foster a learning intention that makes critical thinking self-evident and necessary.

The Cheesy Conclusion: Focus on the Learning, Not Just the Tool

The core is not the tool, but the intention. The skill lies in consciously choosing the right intention—convenience, exploration, deepening, or others—for each (sub)task. This is not a one-time decision, but a continuous process of planning, monitoring, and adjusting. It is the essence of self-regulation, built on metacognitive skills and supported by a solid knowledge base, because without that foundation, there is little to critically assess.

The challenge for education is therefore to foster a learning climate in which students feel the motivation and develop the metacognitive baggage to make those conscious choices. This is where our craftsmanship as education professionals makes the difference. The task is to consciously make time and space for this in our teams, and to shape the conversation about this new/old art of learning together.