Plagiarism and AI Thresholds in Academic Theses: Safeguarding Quality in Distance Education
- OUS Academy in Switzerland
- Jul 31
- 5 min read
This article examines the growing importance of academic integrity in the era of artificial intelligence (AI) and distance learning. A recent policy adopted by a major university sets clear thresholds for plagiarism and AI-generated content in academic theses: less than 10% similarity is acceptable, 10–15% requires evaluation, and above 15% results in failure. Drawing from recent practices and academic studies, this article analyses the implications of such policies for higher education quality, especially for distance learning programmes. The discussion also highlights challenges, possible improvements, and strategies for global universities to ensure credibility and fairness in academic work.
Introduction
The landscape of higher education has changed dramatically in the past decade. Distance learning, online supervision, and digital thesis submissions have become routine in many universities worldwide. Alongside this shift, the rapid growth of AI tools has transformed how students research, write, and submit academic work.
While technology brings opportunities, it also raises concerns about plagiarism and the misuse of AI to generate entire sections of academic theses. Universities increasingly recognise that maintaining academic integrity requires not only sophisticated detection tools but also clear, transparent policies that define acceptable levels of similarity and AI-generated content.
A recent case from a leading university illustrates this well. Under a new policy, PhD theses are screened using advanced detection software, and results are interpreted using specific thresholds:
Less than 10% similarity = Acceptable
10–15% similarity = Needs further evaluation
Above 15% similarity = Fail
This structured approach provides clarity for students, faculty, and examiners. This article examines this policy in depth, explores its implications for distance learning, and offers practical recommendations for universities worldwide.
Literature Review
1. Plagiarism in Higher Education
Plagiarism has been a longstanding issue in academic writing. Studies show that many cases occur not because of deliberate dishonesty but due to misunderstanding citation rules, paraphrasing techniques, or academic conventions. However, intentional plagiarism, especially in high-stakes submissions like PhD theses, remains a serious concern.
2. AI and Academic Integrity
The rise of AI writing tools adds a new dimension. While AI can assist with grammar, translation, and summarisation, it can also produce entire essays or chapters. Academic journals have already reported cases where AI-generated content slipped through peer review processes. Researchers argue that universities need policies defining which uses of AI are legitimate (e.g., editing) and which cross ethical boundaries (e.g., full drafting).
3. Threshold Policies and Detection Tools
Many universities use similarity detection software such as Turnitin, iThenticate, or Drillbit. Thresholds vary, but the <10%, 10–15%, >15% model is gaining acceptance because it balances academic rigour with fairness. This three-tier system avoids harsh penalties for minor or accidental overlaps while addressing serious misconduct.
4. Distance Learning Challenges
In traditional on-campus programmes, supervisors often know students personally and can detect irregularities in writing style or quality. In distance learning, interactions are fewer, increasing reliance on automated tools. Clear plagiarism and AI thresholds thus become essential for fairness and consistency across remote settings.
Methodology
This article adopts a descriptive and analytical approach, combining:
Case Study Analysis: Examining the university that introduced the <10%, 10–15%, >15% policy for PhD theses.
Comparative Review: Analysing similar policies in international universities to identify best practices.
Theoretical Insights: Using concepts from academic integrity research, educational technology, and distance learning pedagogy to interpret findings.
Sources include academic journals, higher education policy reports, and interviews published in education magazines. While specific universities are not named here, the analysis reflects real practices documented in recent studies.
Analysis
The new policy’s three-tier thresholds offer several advantages and raise important questions for implementation.
1. Clarity for Students and Faculty
Students often feel anxious about plagiarism checks because policies can be vague. A fixed threshold removes ambiguity: if the similarity score is below 10%, the thesis proceeds smoothly; between 10–15%, it undergoes review; above 15%, it faces rejection. This transparency helps students plan their writing and supervisors guide them effectively.
2. Balancing Innovation and Integrity
The evaluation zone (10–15%) recognises that not all similarity is unethical. Overlaps may occur in literature reviews, technical terminology, or methodology descriptions. Moreover, limited AI assistance for grammar or language polishing might be acceptable. This middle range allows academic judgment rather than relying solely on software scores.
3. Implications for Distance Learning
In distance education, where faculty-student interactions are mostly online, thesis evaluation often depends heavily on digital tools. Automated similarity checks provide consistency across geographically dispersed students, ensuring equal standards regardless of location.
4. Technology Limitations
Detection software varies in database access and AI-detection accuracy. Some tools may miss plagiarism from restricted journals or paraphrased content. Others may falsely flag common phrases or citations. Therefore, human oversight remains critical, especially in the 10–15% evaluation range.
5. Ethical Use of AI
Students increasingly use AI for idea generation, grammar correction, or formatting. Universities must define acceptable vs. unacceptable uses. For instance, using AI to improve language clarity may be permitted, while generating entire thesis chapters may not. Clear guidelines prevent confusion and ensure fair enforcement.
Findings
From the analysis, several key findings emerge:
Structured Thresholds Improve Fairness: The three-level system avoids over-penalising minor overlaps while addressing serious plagiarism effectively.
Distance Learning Benefits the Most: Automated checks combined with clear rules provide consistency across remote campuses and online programmes.
Training Is Essential: Students and faculty need workshops on academic integrity, correct citation, and responsible AI use.
Technology Alone Is Not Enough: Human judgment remains necessary, especially for borderline cases or evaluating context behind similarity scores.
Policy Transparency Builds Trust: When students know the exact thresholds and consequences, they are more likely to follow academic integrity rules.
Discussion
The combination of plagiarism detection software, AI-content identification, and fixed thresholds represents a pragmatic approach for modern universities. However, successful implementation requires:
Regular Policy Review: As AI tools evolve, detection methods and ethical guidelines must be updated.
Global Benchmarking: International collaboration can help align standards across countries, especially for joint or distance degree programmes.
Student Support Systems: Online tutorials, academic writing centres, and AI-ethics modules can reduce unintentional violations.
Appeal Mechanisms: Students should have the right to contest results if they believe similarity scores misrepresent their work.
Conclusion
The adoption of <10%, 10–15%, >15% thresholds for plagiarism and AI-generated content marks an important step toward preserving academic integrity in the digital era. For distance education, where reliance on technology is higher, such policies ensure consistent, transparent, and fair evaluation of academic theses.
While detection tools provide the technical backbone, true academic quality also depends on student education, ethical clarity, and continuous adaptation to emerging technologies. Universities worldwide can learn from this model to balance innovation with integrity, ensuring that even in the age of AI, academic work remains original, credible, and respected.
References
Gupta, A., Mahaseth, H., & Bajpai, A. (2025). AI Detection in Higher Education: Challenges and Strategies. Academic Integrity Journal.
Toker, S., & Akgun, M. (2024). “Task Complexity and AI-Generated Text: Reducing Academic Misconduct.” Journal of Educational Technology Studies.
Sağlam, T., & Schmid, L. (2025). “Evaluating Automated Plagiarism Detection in Distance Learning Contexts.” International Review of Academic Integrity.
Office of Academic Ethics (2025). Annual Report on Plagiarism and AI in Universities.
Patel, R., & Johnson, L. (2023). Distance Education and Academic Quality Assurance. Higher Education Press.
Comments