AI Use in Higher Education: HFD Study 2026 | Alphabees

The Hochschulforum Digitalisierung (HFD) has published Working Paper No. 91, a comprehensive systematic review of AI use in higher education teaching and learning. The analysis summarizes 15 quantitative studies from 2022 to 2025 and paints a nuanced picture of the current situation at German universities. For decision-makers in education, this assessment provides valuable insights – while also revealing where structured action is needed.

A Transitional Phase with Tensions

Authors Elke Bosse, Klaus Wannemacher, and Maren Lübcke describe the current situation as a challenging transitional phase. Students frequently use AI tools spontaneously, but without deeper reflection on how they work or their ethical implications. Educators experiment selectively with the new possibilities, while institutions develop initial strategic approaches.

This complex landscape exists within a field of tension between two poles: on one hand, the desire for clear rules and regulation; on the other, the goal of integrating AI constructively and with media literacy into university teaching. For education leaders, this means: waiting is not an option. Those who do not actively shape this development risk allowing unreflective usage patterns to become entrenched.

How Students Actually Use AI

The study identifies five central areas of student AI use:

Text work:

Students use AI tools to write, revise, or summarize texts.

Learning support:

The focus here is on clarifying comprehension questions and gaining an overview of complex topics.

Programming and data analysis:

Particularly in technical fields, AI tools serve as support for code creation and data analysis.

Presentation and design:

Visual preparation of content is facilitated by AI-powered tools.

Study and self-organization:

Planning and structuring tasks are increasingly delegated to AI systems.

The study reaches a sobering conclusion: usage aims less at conceptual learning than at relief and efficiency. Students seek quick answers, not deeper understanding. Furthermore, knowledge about technical functionality and ethical questions appears to be limited.

This presents universities with a central task: the existing willingness to use AI must be channeled into reflective, learning-enhancing directions. It is not enough to ban AI tools or allow them without regulation. What is needed are pedagogical concepts that establish AI as a tool for deeper learning.

The Perspective of Educators

Unlike with students, data on AI use by educators is considerably thinner. However, the available findings show clear priorities: educators primarily use AI tools for planning and preparing courses. Pedagogical design also benefits from the new possibilities.

What is often missing in practice is the integration of AI support directly into students' learning processes. Here lies untapped potential: a systematically embedded AI tutor could bridge precisely the gap between students' desire for relief and educators' expectations for reflective learning.

Methodological Limitations and a Dynamic Field

The authors of the HFD study themselves point to important limitations. Since each of the 15 analyzed studies uses its own methodological design, comparability of results is limited. The dynamics of the research field exacerbate this problem: what is collected about AI use today may already be outdated tomorrow.

Notably, specific AI tools barely feature in the analyzed studies. The question of which specific applications students and educators use remains largely unanswered. For universities that need to make concrete implementation decisions, this represents a knowledge gap.

However, these methodological limitations do not change the core finding: the use of AI in higher education is a reality – and it currently proceeds largely unguided.

From Assessment to Strategic Implementation

The HFD study provides a solid foundation for strategic decisions. It shows that superficial usage patterns dominate and that neither students nor educators currently possess sufficient AI competencies. At the same time, it becomes clear: the willingness to use AI exists, but institutional frameworks lag behind.

For universities and continuing education providers, this points to concrete areas for action:

• Development of clear guidelines for AI use in teaching and learning
• Training educators in AI-supported pedagogy
• Integration of AI tools that promote reflective learning rather than just increasing efficiency
• Building infrastructure that enables privacy-compliant and context-specific AI use

The Alphabees AI Tutor for Moodle addresses precisely these requirements. As a learning companion integrated directly into existing Moodle courses, it supports students around the clock – not as a shortcut to quick results, but as a tool for deeper understanding. Educators retain control over pedagogical integration, while institutions benefit from a privacy-compliant solution tailored to the respective course context.

The HFD study makes clear: the question is no longer whether AI plays a role in higher education, but how that role is shaped. Those who set the right course now can use the transitional phase to make reflective AI use the standard – thereby sustainably improving both teaching quality and learning outcomes.

Frequently Asked Questions