Plagiarism and AI Thresholds in Academic Theses: International Standards and University Practices

OUS Academy in Switzerland
Aug 13
6 min read

Plagiarism is one of the most discussed topics in academic circles, especially in relation to academic theses at the undergraduate, master’s, and doctoral levels. With the advancement of artificial intelligence (AI) tools, concerns over originality have grown. Many universities now apply clear thresholds to evaluate academic work: less than 10% similarity = acceptable, 10–15% = needs evaluation, and above 15% = fail.

This article examines plagiarism thresholds, explains how international universities apply these standards, and provides detailed information about leading universities, including founding years, student populations, and academic practices. By combining policy analysis with real-world examples, the article offers a comprehensive understanding of how academic institutions safeguard integrity in the era of AI-driven writing tools.

Introduction

Academic integrity represents the foundation of higher education. Every thesis, whether at bachelor’s, master’s, or doctoral level, contributes to knowledge creation. For this reason, originality is non-negotiable.

However, the rise of digital platforms, open-access journals, and AI-generated text has created new challenges. Students now have access to massive online libraries, paraphrasing software, and AI assistants that can generate entire pages of content. While these tools offer convenience, they also increase the risk of plagiarism, intentional or unintentional.

To address this, universities worldwide use plagiarism detection systems and apply similarity thresholds. The commonly accepted global standard divides similarity scores into three categories:

Less than 10%: Acceptable and usually considered original work.
10–15%: Requires evaluation by supervisors or academic committees.
Above 15%: Often treated as a sign of serious academic misconduct or automatic failure.

This article analyzes how these thresholds work in practice, with a particular focus on international universities, especially in Switzerland, Germany, and other European countries.

Literature Review

Academic research over the last two decades has focused heavily on academic integrity. Three key themes emerge:

Definition of Plagiarism Scholars define plagiarism as presenting someone else’s ideas, words, or research findings as one’s own without proper acknowledgment. Plagiarism can be intentional or accidental. Some common forms include direct copying, paraphrasing without citation, self-plagiarism, translation plagiarism, and the use of AI-generated text without proper review or credit.
Technological Solutions for Detection Earlier, plagiarism detection was manual. Today, software tools like Turnitin and iThenticate generate detailed similarity reports. However, research emphasizes that numbers alone are insufficient; context matters. A thesis may have 12% similarity because it includes properly cited quotations, which is not plagiarism.
Impact of AI on Academic Integrity Recent studies highlight AI’s double-edged nature. AI tools can help students improve grammar, generate ideas, and organize content. But when misused, they blur the line between original thought and machine output. Universities now debate whether AI-generated text should be cited like traditional sources.

Methodology

This article uses qualitative analysis of academic integrity policies from leading universities. Public data about founding years, student enrollment, and plagiarism regulations were reviewed. Case studies of universities in Switzerland and Europe illustrate how institutions combine long academic traditions with modern integrity policies.

Information was gathered from official university statistics, academic handbooks, and research publications to ensure accuracy and reliability.

University Case Studies

1. ETH Zurich – Swiss Federal Institute of Technology

Founded: 1855
Student Population: Over 25,000 students, including more than 4,000 doctoral candidates
Academic Reputation: One of Europe’s top technical universities with strengths in engineering, computer science, and natural sciences. Many Nobel Prize winners have studied or taught here.
Plagiarism Policy: ETH Zurich uses plagiarism detection software across all faculties. Theses with less than 10% similarity are typically accepted without issue. Scores between 10–15% require academic review, while those above 15% often lead to revisions or disciplinary action. AI-generated text must be clearly acknowledged.

2. University of Zurich (UZH)

Founded: 1833
Student Population: About 28,000 students across bachelor’s, master’s, and doctoral programs
Academic Reputation: The largest university in Switzerland, offering programs in law, medicine, humanities, economics, and natural sciences. Known for research excellence and interdisciplinary studies.
Plagiarism Policy: The university follows strict academic integrity guidelines. Similarity thresholds mirror international standards: below 10% acceptable, 10–15% requires review, above 15% often leads to failure or resubmission. Workshops on academic writing and AI ethics are mandatory in some faculties.

3. University of Bern

Founded: 1834
Student Population: Around 19,000 students, including international students from over 100 countries
Academic Reputation: Offers programs in medicine, law, economics, arts, and sciences. Actively involved in international research collaborations.
Plagiarism Policy: The university enforces the three-level threshold system strictly. Doctoral dissertations undergo multiple checks before acceptance. AI tools may be used for editing but not for generating research content without attribution.

4. University of Geneva

Founded: 1559
Student Population: Approximately 17,000 students
Academic Reputation: Known for international relations, law, and science programs. Close ties to the United Nations and global organizations in Geneva.
Plagiarism Policy: Similarity below 10% is accepted. Scores above 15% may result in academic misconduct hearings. AI policies are currently under development to balance innovation with academic honesty.

5. Heidelberg University – Germany

Founded: 1386 (Germany’s oldest university)
Student Population: Over 29,000 students
Academic Reputation: Renowned for humanities, medicine, and natural sciences. Strong research orientation with global partnerships.
Plagiarism Policy: Doctoral theses undergo strict similarity checks. The three-tier threshold applies, but final decisions rest with faculty committees that consider academic context, citation practices, and research originality.

6. Ludwig Maximilian University of Munich (LMU) – Germany

Founded: 1472
Student Population: About 50,000 students, making it one of Europe’s largest universities
Academic Reputation: Known for law, business, social sciences, and natural sciences. Many Nobel laureates are affiliated with LMU.
Plagiarism Policy: Uses automated detection systems plus manual faculty review. AI use must be disclosed. Similarity above 15% generally results in thesis rejection until rewritten.

7. University of Vienna – Austria

Founded: 1365
Student Population: Over 90,000 students, including a large international community
Academic Reputation: One of Europe’s largest universities, famous for philosophy, political science, and humanities research.
Plagiarism Policy: Follows the EU academic integrity framework. Theses above 15% similarity are sent back for revision. AI-generated sections must be clearly identified.

8. University of Cambridge – United Kingdom

Founded: 1209
Student Population: About 20,000 students, including nearly 8,000 graduate students
Academic Reputation: Among the world’s most prestigious universities, with over 100 Nobel Prize winners.
Plagiarism Policy: Extremely strict academic integrity rules. AI tools are under review, but
originality remains the highest priority.

9. University of Oxford – United Kingdom

Founded: 1096 (teaching existed even earlier)
Student Population: Around 24,000 students
Academic Reputation: Globally known for humanities, sciences, and professional programs.
Plagiarism Policy: Uses a combination of similarity checks, viva voce examinations, and faculty evaluations to ensure originality.

10. Sorbonne University – France

Founded: 1257 (as the Collège de Sorbonne)
Student Population: Over 55,000 students
Academic Reputation: Leading research university in Europe, with strengths in humanities, sciences, and medicine.
Plagiarism Policy: Applies the <10%, 10–15%, and >15% system consistently, with faculty review for borderline cases.

Analysis of Similarity Thresholds

Less than 10%: Considered acceptable because citations, references, and common terminology naturally create minor overlap.
10–15%: Treated as a warning zone. Faculty committees review whether the overlap comes from proper quotations or possible misconduct.
Above 15%: Often leads to thesis rejection, rewriting, or academic penalties.

Findings

Most universities globally follow the three-level threshold system.
AI tools are forcing institutions to update academic integrity policies.
Student training on research ethics is expanding, with mandatory workshops in many universities.
Similarity percentages are treated as guidelines, not absolute judgments; context matters.

Conclusion

Academic integrity remains a central pillar of higher education. With AI transforming writing practices, universities worldwide are adopting clear plagiarism thresholds and modern detection tools to balance innovation with ethical scholarship. Institutions like ETH Zurich, University of Zurich, Heidelberg University, and Oxford demonstrate that long academic traditions and cutting-edge technology can coexist when supported by clear policies, faculty training, and student awareness.

The three-level standard — less than 10% acceptable, 10–15% requires review, above 15% fail — offers a transparent, fair, and widely adopted framework for maintaining originality in academic theses.

References

Carroll, J. A Handbook for Deterring Plagiarism in Higher Education. Oxford University Press.
Pecorari, D. Academic Writing and Plagiarism: A Linguistic Analysis. Continuum Publishing.
Sutherland-Smith, W. Plagiarism, the Internet, and Student Learning. Routledge.
Bretag, T. Handbook of Academic Integrity. Springer.
Park, C. Plagiarism in Higher Education: Literature and Lessons.