Agentic Learning – The Learning Agency of the Future?

What differentiates these individuals from intrinsically motivated self-learners?

Thorsten Zylowski: Self-learners typically learn independently and out of intrinsic motivation, often outside of formal learning structures. Agentic learning goes a step further: it focuses not only on motivation but also on the conscious management of the entire learning process.

Agentic learners actively set goals, choose strategies, use resources in a targeted manner, and reflect on their progress. Furthermore, agentic learning often takes place in designed learning environments that support this self-direction, for example through choices, feedback systems, or digital tools. While self-directed learning describes more of an individual characteristic, agentic learning is also a didactic design principle.

The term "agentic learning" can apply to both humans and AI. So, where’s the difference?

Thorsten Zylowski: The distinction is primarily based on the nature of Agency. Humans have intentional Agency: They develop their own goals, values, and meanings and can relate their learning to personal interests or life contexts. AI systems, on the other hand, have functional or simulated Agency. They can plan tasks, execute steps, and make decisions within a given framework, but they do not pursue their own goals or motivations. In the context of agentic learning, this means:

•    Humans are the true drivers of learning agency.

•    AI can support this agency, for example as a coach, research assistant, or reflection partner.

•    The challenge is to design systems in such a way that they strengthen autonomy rather than replace it.

What are the most common use cases for agentic learning today?

Thorsten Zylowski: There are already several practical fields where elements of agentic learning are being implemented:

•    AI-powered learning assistance: Systems that support learners in conducting research, organizing content, or understanding complex topics.

•    Adaptive learning platforms: Learning environments in which users can choose their own learning path and support is dynamically adapted.

•    Project- and problem-based learning: Learners define their own questions and use digital tools and AI to develop solutions.

•    Reflection and learning journals with AI support: AI helps learners reflect on their progress and plan next steps.

In many cases, agentic learning is particularly evident when learners actively use technologies to manage their own learning processes.

What potential do you see for agentic learning in the future?

Thorsten Zylowski: Agentic learning has the potential to make learning more personalized, continuous, and context-driven. AI systems could serve as personal learning agents that support learners in setting goals, planning, and reflecting. They would do this without taking control of the learning process. In the long term, learning environments could develop in which humans and AI work together in a complementary way: humans define goals, perspectives, and meanings, while AI helps structure information, identify options, and provide feedback.

Therefore, the primary potential of this approach isn't just in more efficient learning, but in the development of learning competence and self-efficacy—skills that are becoming increasingly important in an increasingly complex and technology-driven world.