Electronic examinations (e-examinations for short, sometimes also online examinations) have played an important role in higher education for some time. A distinction is made between electronic face-to-face examinations, i.e. those that are taken on a computer but still in person at the university, and electronic remote examinations, in which the examinees are at different locations (usually outside the university, e.g. in their own home). Examinations can also take place synchronously or asynchronously, supervised or unsupervised and can be conducted in various formats, including written, oral and practical. The term “e-exams” therefore covers a wide range of different exams, depending on the degree program, subject and conditions at the university.
While distance learning universities offered digital examination options early on in order to enable their students to take examinations as flexibly as possible, most universities only began to take a closer look at the opportunities and challenges of these examinations during the COVID-19 pandemic. Due to the contact restrictions, hygiene requirements and travel restrictions in place at the time, even universities that were designed for purely face-to-face teaching had to provide alternatives from 2020. In particular, it was often no longer possible to offer the traditional written exam (a mostly handwritten exam, which was supervised and took place in a fixed time slot at the universities) in person. In many federal states, legal bases for e-distance examinations were therefore created (in Bavaria, for example, the BayFEV), which enable universities to conduct examinations remotely – supervised by camera and microphone. Since the use of a computer and the Internet (for supervision and for the transmission of examination details and examination performance) is a prerequisite and at the same time makes it more difficult to detect certain acts of cheating (e.g. the use of a cheat sheet), remote e-examinations are helping to rethink the current examination culture. On the one hand, the computer can be used to create more modern, competence-oriented examinations that are adapted to later professional life (e.g. through the use of programming tasks, interactive environments or videos); on the other hand, it offers the opportunity to make the examination itself more inclusive and therefore more equal in terms of opportunities. For example, an individual disadvantage can sometimes be better compensated for with electronic examinations than is the case with handwritten examinations (e.g. through individual contrast settings on the screen or a Braille keyboard). In many cases, the open-book format, which is particularly widespread in the Anglo-American culture, is also used, where students can access teaching materials and other aids during their exam. This reduces the need for supervision and therefore the need for technical supervision measures that invade fundamental rights.
The main practical challenges lie in setting up the infrastructure and training all those involved. On the legal side, complex trade-offs between the fundamental rights of different fundamental rights holders are required. In addition to the principle of equal opportunities, the students’ right to examination, the academic freedom of the lecturers and, where applicable, the protection of life and health, the right to informational self-determination and the so-called basic computer right (basic right to guarantee the confidentiality and integrity of information technology systems) must be taken into account. The latter guarantees system protection of privately used IT devices from state access and thus sets limits on the mandatory installation of examination software, for example. If students are in their own homes during the video-supervised examination, the right to inviolability of the home may also be affected.
With the availability of ChatGPT in fall 2022, the university examination system has continued to change. A large number of AI-based assistants are now available to everyone, which can not only write texts, but also calculate, program or generate images and videos in response to the simplest instructions. In the course of this development, the value of unsupervised examination work, such as seminar papers or theses, has been discussed in particular. This not only raises the question of whether the use of generative AI should be permitted, but also sheds light on which skills are to be demonstrated by such an examination, which standards apply and how it can be verified that the performance was essentially provided by the student.
A uniform approach to generative AI systems is not yet apparent in the university examination system. The first court decisions on this (albeit in interim legal protection) (VG Munich of 28.11.2023 – M 3 E 23.4371 and of 08.05.2024 – M 3 E 24.1136) show that there are also problems on the part of the judiciary in dealing with this phenomenon using the familiar legal standards. Challenges arise, particularly with regard to the independence of a performance and what requirements universities have to meet if they want to prove that an examination performance was created illegally (as it was not permitted in this case) by or with the help of (generative) AI. If the use of AI cannot be proven or cannot be proven in accordance with the law, i.e. for example on the basis of a corresponding legal basis for the use of so-called AI detectors, a ban cannot be enforced and would ultimately be arbitrary. The future cannot therefore lie in a ban on this technology, but must integrate it into the audit system, for example in the form of permissible co-creation.
The current and foreseeable developments in university examinations are also accompanied by various data protection and examination law challenges. At the same time, new technologies not only offer the opportunity to adapt the examination system to the changing requirements of the world of work, but also force us (as in the case of generative AI) to rethink our previous examination formats and the value of examinations.
Comparability with analogue phenomena
Parallels can be drawn to a certain extent with previous university examinations. In traditional (analog) open-book exams, students can access teaching materials to solve complex problems, which enables a practical and reality-based exam. This can be extended to electronic remote exams and the use of AI applications can also be permitted. This involves accessing not only books but also intelligent tools, which places an even greater focus on problem solving and creative approaches (generation and processing of large amounts of data, networking and data integration). The possibilities for designing e-exams are much more diverse than analog exams, and the content and sequence of tasks can be varied more easily and adapted more quickly (increased changeability, high speed).
Supervision in the context of e-(remote) examinations is also initially modeled on the analog counterpart. However, supervision and control measures are often implemented digitally in the case of e-examinations and in some cases are already carried out by proctoring software to ensure that the examination conditions are fair and secure (automation, simplified identification). While face-to-face exams require physical presence at a central location, e-exams offer greater flexibility thanks to video proctoring (electrical/optical transmission and processing), as they can be conducted regardless of location. This can facilitate access to examinations and better meet the needs of a diverse student body (ubiquitous availability). Furthermore, the electronically processed examination results can be transmitted directly (and pseudonymized if necessary) to the examiners. If assessments are to be obtained from several examiners, this can be done simultaneously – in contrast to analog processing – so that the correction process can be made more efficient (lossless duplication, networking and data integration).
Nevertheless, e-audits and also audits that are (or can be) processed with the help of AI have decisive differences to the familiar analog audits, which stand in the way of a transfer of all (audit) legal considerations. On the one hand, the necessary data processing leads to new and more complex conflicts of fundamental rights. On the other hand, the universities’ ability to influence the examinations is reduced. Certain factors can no longer be unilaterally controlled by them, e.g. universities can no longer influence the specific examination environment in the case of remote e-examinations. Instead, this is the responsibility of the examinees themselves. The handling of universities with regard to the possible use of AI in unsupervised examinations has also been limited to date; there is often a lack of clear guidelines on what is permitted and what is prohibited. It is also unclear whether and how a violation of an AI ban can be proven and enforced.
Social relevance
The introduction of digital technologies in the examination system is not limited to the mere electrification of examinations, but is leading to far-reaching changes in the education and higher education system. The AI applications currently available are already showing the limits of the tried-and-tested examination formats and require a reconsideration of the university examination system. For example, it must be questioned which skills and knowledge are relevant and required in the world of work and must be taught by universities. How can and should graduates be prepared for the labor market of the future, in which digital technologies are omnipresent and the use of AI-supported assistants is not only possible, but even required? What is the value of a text written by a human compared to a purely AI-generated text or one that results from the co-creation of humans and machines?
The digital transformation is also leading to certain shifts in the competence structure between universities and students. Whereas the framework conditions for analog examinations were previously determined entirely by the universities, digital examinations increasingly require cooperation. However, these changes also offer the opportunity to meet the challenges with courage and a creative spirit, to gain experience in dealing with the new technologies and thus to further develop the examination system together.
Further links and literature
- Bayerisches Kompetenzzentrum für Fernprüfungen.
- Bayerisches Staatsinstitut für Hochschulplanung und Hochschulforschung, Evaluation der Bayerischen Fernprüfungserprobungsverordnung (BayFEV) – Abschlussbericht Juni 2024.
- Braegelmann, T. (2024). Zuhilfenahme Künstlicher Intelligenz bei der Erstellung von Texten für die Universität. RDi, 188-192.
- Heckmann, D./Rachut, S. (2024). Warum KI-Einsatz bei Klausuren erlaubt sein sollte. Deutschlandfunk Nova v. 21. Juni 2024
- Heckmann, D./Rachut, S., (2024). Rechtssichere Hochschulprüfungen mit und trotz generativer KI. In: Ordnung der Wissenschaft 02, 85–100.
- Heckmann, D./Rachut, S. (2023). E-Klausur und Elektronische Fernprüfung.
- Rachut, S. (2024). Grundrechtsverwirklichung in digitalen Kontexten.
- Rachut, S. (2024). Kein Zugang zum Masterstudium wegen Vorlage eines mittels KI erstellten Essays. Anmerkung zu VG München, B. v. 28.11.2023 – M 3 E 23.4371, NJW 2024, 1052–1057.
- Rachut, S.(2023). E-Klausur und elektronische Fernprüfung: Technologischer Fortschritt und Prüfungskulturwandel im Spiegel des Rechts. Ein Werkstattbericht. In: OdW 02, 89–98.
- Rachut, S. (2023). Exmatrikulation aufgrund schwerwiegender Täuschung im Rahmen einer elektronischen Fernprüfung – Anmerkung zu VG Berlin 12. Kammer, Urteil vom 06.02.2023 – 12 K 52/22, jurisPR-ITR 19/2023, Anm. 4.
- Rachut, S. (2021). Hochschulprüfungen als Fernklausur. Anm. zu VG Frankfurt (Oder), Beschl. v. 11.05.2021 – 1 L124/21, COVuR 2021, 488–493.