New Roles for Students – From Consumer to Active Curator of Their Own Learning Through Generative AI

Students are increasingly using AI not just as an aid, but as a replacement for their own thinking – with far-reaching consequences. However, instead of bans, we need strategies. This article shows how generative AI can be used to enhance learning – through four central roles that are reshaping education. 

GenAI Impedes the Generation Effect

When teachers speak out to the public in alarm, we should listen carefully. In an article in the Frankfurter Rundschau, a German teacher urgently warns of a “risky trend”: More and more students are using generative AI tools like ChatGPT to complete homework, presentations, or even exam assignments in a fully automated manner. [1] The concern: technological development is outpacing school reality – and gradually undermining pedagogical control.

A US teacher expresses similar criticism: She reports on a situation where she finds it increasingly difficult to distinguish between genuine student work and AI-generated output. The ability to clearly structure thoughts, write independently, or critically question complex content – all of this is declining among her students. [2] A central reason for this lies in the so-called Generation Effect: learners remember and understand content more sustainably when they actively formulate it themselves, rather than merely consuming it. Those who instead rely on generative AI to obtain ready-made texts bypass this cognitive process – and thereby risk blocking deeper understanding and long-term learning. [3] The message from both voices is clear: artificial intelligence threatens the traditional educational canon – and encounters an education system and learners who are hardly prepared for it.

Using GenAI as a Productive Learning Tool

However understandable these warnings may be, they often fall short. Because generative AI (GenAI) is far from being just a risk – it is also an opportunity. An opportunity to make learning processes more individualized, to promote digital competencies, to relieve teachers’ workload, and to make education inclusive and efficient in new ways. The difference lies – as so often – in how it is used.

The study by Seco, Grösser and Pedrosa (2025) provides an overview of the status quo and potential of generative AI in higher education. Jeon and Lee (2022) have identified four core functions in which GenAI systems like ChatGPT can be effectively deployed: as interlocutor, content provider, teaching assistant, and evaluator. These roles were not only theoretically derived, but also examined in a practical test with language instructors – with exciting implications for designing future university teaching. The four roles are described in detail below:

1. Interlocutor – GenAI as a Dialogue-Based Learning Partner

In its role as interlocutor, GenAI becomes a vehicle for realistic interaction. Students can engage in simulated conversations with AI-powered chatbots – for example, with fictional customers, employers, supervisors, or colleagues. Particularly in the area of language and communication training, this function offers enormous potential: GenAI can conduct dialogues naturally, grasp context, and answer individual follow-up questions. This is especially advantageous in foreign language degree programs or professional training where confident expression, technical terminology, and cultural communication patterns are crucial. Studies show that GenAI-based language dialogues not only promote language practice, but also improve confidence and speaking competence (Kim, 2019).

2. Content Provider – GenAI as a Dynamic Content Engine

In the role of content provider, AI supports educators in creating, processing, and personalizing teaching materials. GenAI can automatically generate texts, assignments, summaries, quiz questions, or even graphic explanations – tailored to difficulty level, topic, or target group. Audio or video formats can also be generated or adapted (with limitations and usually at a cost). This function is particularly relevant in higher education, where individual in-depth materials or adaptive content are often needed. For example, GenAI can generate simplified explanations for weaker students, while creating more advanced case studies or transfer tasks for advanced learners. In doing so, GenAI does not function as a replacement for subject matter expertise, but as an extension: it provides drafts, ideas, and variations that students can further develop and critically examine. The quality – especially in the text domain – is now competitive with human-created content, although limitations exist in terms of originality, depth, and multimodality (e.g., video, audio) (Kim et al., 2020).

3. Teaching Assistant – AI as an Intelligent Tutoring System

As a teaching assistant, GenAI functions as a so-called Intelligent Tutoring System that accompanies students in their learning process. This involves not just providing information, but adaptive support: the AI analyzes learning behavior, identifies knowledge gaps, provides individualized feedback, and suggests next learning steps. For example, it can give students targeted hints after incorrectly answered tasks, suggest alternative solution methods, or offer additional materials. This function enables close-knit, data-driven supervision – particularly in large courses or digital learning settings where individual support would otherwise be difficult to implement. Studies show that such systems can bring significant advantages especially for lower-performing or inexperienced learners – for instance through targeted scaffolding, i.e., the gradual guidance toward independence (Hooshyar et al., 2015). However, the prerequisite is intelligent pedagogical integration: AI should support, not replace. Excessive automation can lead to passivity and dependency.

4. Evaluator – GenAI as Support for Assessment and Examinations

In the role of evaluator, GenAI assists with creating, conducting, and grading assessments. It can generate multiple-choice questions, formulate exam tasks, provide preliminary assessment of essays, or automate feedback. Particularly in large courses, this offers enormous time savings and a certain standardization. One example is the automated scoring of essays or short answers, where GenAI examines content based on criteria such as coherence, argumentative logic, or grammar. Initial studies show that agreement with human assessments already reaches over 95% in some cases (Gierl et al., 2014). Nevertheless, skepticism remains warranted: assessment standards must be transparent, comprehensible, and pedagogically sound – this cannot be achieved by algorithms alone.

Final Thoughts

Those who respond to the challenges of generative AI only with bans miss its educational policy opportunity – because only through targeted pedagogical integration, critical reflection, and new roles for learners can technological disruption become genuine progress for teaching and learning.

Sources

[1] https://www.fr.de/panorama/macht-das-leben-zur-hoelle-lehrerin-kritisiert-riskanten-trend-an-deutschen-schulen-zr-93738219.html

[2] https://www.fr.de/panorama/us-lehrerin-schlaegt-alarm-ki-veraendert-das-schreiben-der-schueler-grundlegend-zr-92756678.html

[3] https://www.linkedin.com/feed/update/urn:li:activity:7330134708107755521/

Gierl, M.J., Latifi, S., Lai, H., Boulais, A.P., Champlain, A. (2014). Automated essay scoring and the future of educational assessment in medical education. Medical Education, 48(10), 950-962. https://doi.org/10.1111/medu.12517

Hooshyar, D., Ahmad, R.B., Yousefi, M., Yusop, F.D., Horng, S.J. (2015). A flowchart-based intelligent tutoring system for improving problem-solving skills of novice programmers. Journal of Computer Assisted Learning, 31, 345-361. https://doi.org/10.1111/jcal.12099

Kim, N.Y. (2019). A Study on the Use of Artificial Intelligence Chatbots for Improving English Grammar Skills. Journal of Digital Convergence, 17(8), 37-46. https://doi.org/10.14400/JDC.2019.17.8.037

Kim, J., Shin, S., Bae, K., Oh, S., Park, E., del Pobil, A. P. (2020). Can AI be a content generator? Effects of content generators and information delivery methods on the psychology of content consumers. Telematics and Informatics, 55, 101452. https://doi.org/10.1016/j.tele.2020.101452

Lee, S., & Jeon, J. (2022). Visualizing a disembodied agent: young EFL learners’ perceptions of voice-controlled conversational agents as language partners. Computer Assisted Language Learning, 37(5–6), 1048–1073. https://doi.org/10.1080/09588221.2022.2067182

Seco, D., Grösser, S. N., & Pedrosa, A. M. (2025). Use of Generative Artificial Intelligence tools in higher education environments. Multidisciplinary Journal for Education, Social and Technological Sciences, 12(1), 156–175. https://doi.org/10.4995/muse.2025.23623