The AIffectiveness project is a collaborative research initiative investigating the effectiveness of Artificial Intelligence in higher education. The project is being conducted by bidt (Bavarian Research Institute for Digital Transformation), OneTutor, and ten Bavarian universities and colleges, with bidt taking the scientific lead.
Project Start: April 2025
Duration: 3 years
Core Goal: To comprehensively investigate the effectiveness and potential of AI tutors in higher education, establishing long-term, evidence-based insights for integrating AI into teaching.
The Central Hypothesis: "Effectiveness" is multidimensional. The impact of AI tutors depends heavily on the specific context: the subject matter, lecture size, the student's current stage of studies, the mathematical component, and whether the course focuses primarily on knowledge transmission or critical argumentation.
The research is being led by bidt
Participating Institutions
Technical University of Munich (TUM)
Ludwig-Maximilians-Universität München (LMU)
Virtuelle Hochschule Bayern (vhb)
Technical University of Nuremberg (UTN)
Rosenheim Technical University of Applied Sciences (TH Rosenheim)
University of Augsburg
Ostbayerische Technische Hochschule Amberg-Weiden (OTH)
University of Bayreuth
Weihenstephan-Triesdorf University of Applied Sciences (HSWT)
Aschaffenburg University of Applied Sciences (TH Aschaffenburg)
You can find a more detailed project description (DE) here:
📊 Initial Research Results:
📰 Press Releases:
FAQ
What is the goal of the research?
The aim is to scientifically analyze whether and how the use of OneTutor delivers measurable benefits. Specifically, the project looks at how it:
Improves learning success,
Deepens subject understanding,
Increves engagement and enjoyment in learning,
Reduces student dropout rates.
Additionally, the project aims to develop standardized methods for evaluating AI-supported teaching.
How is learning success measured?
The bidt is developing comprehensive questionnaires and analytical procedures that combine platform usage data with survey results. This allows researchers to analyze how factors like motivation, comprehension, or trust in AI affect learning outcomes.
⚠️ Note on Privacy: Correlating individual interaction data with final exam grades is not possible. To honor OneTutor's pseudonymization promise, students use the platform in a protected environment. Individual interactions cannot be traced back to specific people or their grades. However, aggregate course-level data can be compared to previous semesters to identify general shifts in performance.
What does participation mean for the universities?
Free Course Budgets: Participating universities receive a budget for OneTutor courses, allowing them to use the platform for free within the scope of the project.
Student Surveys: Students in these courses receive pseudonymized questionnaires that serve as the foundation for the scientific study.
Faculty Involvement: Faculty surveys are optional but highly encouraged.
Staff Funding: Institutions have the opportunity to integrate a research assistant into the project via the allocated research budget.
Which questionnaires are used?
Data collection relies on several distinct surveys:
In-OneTutor Surveys (Main Instruments):
Initial Survey: Displayed to active users after a set number of interactions early in the semester. (A shortened version is shown if a student is already taking it in another course).
Follow-up Survey: Conducted at the end of the semester, displayed only to those who completed the initial survey after a defined time buffer.
Note: Completing these is mandatory to continue using the tool, though "No answer" is an option for every question.
External Student Survey: An optional survey distributed via a link from lecturers to reach non-users.
Lecturer Survey: An optional questionnaire designed to capture the perspective and experiences of the teaching staff.
The content of the In-OneTutor questionnaires can be viewed here.
When will the questionnaires be published?
Release windows are tracked by Calendar Weeks (CW) indicating when surveys are unlocked. Because they trigger based on active tool usage, individual students may see them at different times. Once unlocked, the questionnaires remain permanently active.
What data do participating lecturers receive?
The primary output of the project will be cross-institutional scientific publications by bidt to account for different teaching environments and fields of study.
Course-level results will not be publicly published.
Upon request, lecturers can obtain access to the pseudonymized raw questionnaire data collected within their own specific courses.
What is the legal basis?
A Cooperation and Grant Agreement regulates the scientific framework between bidt and the participating universities.
Data Processing Agreements (DPA / Art. 28 GDPR) are in place between OneTutor and the universities to ensure compliant data handling.
Questionnaires are collected pseudonymously by OneTutor and forwarded to bidt for analysis.
Usage and survey data (e.g., pseudonymous chat, quiz logs, survey inputs) are technically recorded by OneTutor and utilized strictly for the project's research purposes.
All data processing is pseudonymized → individual students cannot be identified.
Scientific findings are published exclusively in an aggregated and anonymized format.
How is data protection guaranteed?
All partners have signed binding cooperation agreements and commit to full compliance with the General Data Protection Regulation (GDPR), the German Federal Data Protection Act (BDSG), and state data protection laws.
If necessary, additional data protection agreements will be concluded in accordance with Articles 26 or 28 of the GDPR.
The data will be processed exclusively:
for a specific purpose (for research and teaching),
in a pseudonymized form (students cannot be identified),
for a limited period of time (only for the duration of the project).