Current status of Turnitin at Japanese univs: A multi-pronged method for academic honesty

The explosive proliferation of generative AI has brought unprecedented transformation and challenges to educational settings. Both the process by which students write reports and papers, and the criteria by which instructors evaluate them, are being fundamentally questioned. Amidst this tumultuous situation, Turnitin, a globally utilized plagiarism detection service, is evolving its role from a mere "misconduct detection tool" into an "educational support platform that fosters academic integrity."

What is Turnitin? Its Functions and the Background of its Spread in Japan

Turnitin is a cloud-based service provided by an EduTech company established in the United States in 1998. Its basic function is to compare the content of reports and papers submitted by students with a vast amount of information on the internet (over 70 billion pages), academic literature databases, and past submissions within the institution, displaying the similarity as a percentage. This makes it possible to detect both unintentional citation of others' writings and malicious plagiarism.

According to a representative from iJapan Corporation, the domestic distributor, the number of institutions that have adopted Turnitin in Japan already exceeded 80 by 2019 and is currently still increasing. There are two major factors behind this widespread adoption.

First, "risk prevention." As symbolized by the STAP cell controversy in 2014, academic misconduct is a serious matter that not only damages an individual's career but also significantly undermines the social credibility of the entire educational institution. Turnitin is also utilized as "a useful tool to assist with the final check" for not only students but also faculty and researchers to review their own writings. This can be described as a systematic measure that complements the academic world, which largely operates on the assumption of good faith, to prevent the occurrence of misconduct.

Second, "improvement in the quality of feedback." As its name "Feedback Studio" suggests, Turnitin offers a variety of functions beyond just plagiarism checking. Instructors can not only insert text comments at any point in a submission but also efficiently provide feedback by dragging and dropping pre-registered standardized comments (QuickMark). Furthermore, it integrates voice comment functionality (Voice Comments) and an English grammar checking engine (e-rater) developed by ETS, the organization that administers TOEFL, making it a highly beneficial tool, especially for English language instructors. In this way, Turnitin enhances its value as an integrated platform for reducing the burden on instructors and providing students with higher quality feedback.

Case Studies of Major Universities and Their Diverse Usage Methods

The introduction of Turnitin in Japanese universities varies significantly in form and purpose, depending on each institution's educational policies and historical background. Below, we introduce examples from representative universities.

• Waseda University: A pioneer that was the first in Japan to extensively implement Turnitin Feedback Studio across the entire university. The introduction semi-automated plagiarism checking for faculty, improving the accuracy and efficiency of checks. However, the AI-generated text detection feature was discontinued as of December 2023, and there are currently no plans for its reintroduction. This is presumed to be a decision based on concerns about the accuracy of AI detection technology and differences in educational approaches.

• The University of Tokyo: Is addressing generative AI with a very cautious and educational perspective. From March to April 2023, it published guidelines for students and faculty, cautioning instructors not to "over-rely on AI detection tools." Detection results are used as clues to understand students' learning processes, and the final evaluation policy emphasizes human review by instructors and dialogue with students.

• Keio University: Does not uniformly prohibit the use of generative AI, but instead adopts a policy that encourages its appropriate utilization. The university has formulated guidelines for the use of generative AI, recommending its application for improving the efficiency of learning, research, and work, provided that users correctly understand the associated risks (information security, copyright, etc.). However, explicit permission from the instructor in charge is required for its use, and users are asked to clearly indicate any parts where AI was utilized.

• Ritsumeikan University: Began partial introduction in the 2017 academic year and embarked on a university-wide implementation starting from the Fall semester of the 2023 academic year. The purpose of its introduction extends beyond plagiarism prevention; it is also utilized from the perspective of educational management, for example, by using it as data to measure students' learning achievement.

• Chuo University: Introduced Turnitin Feedback Studio across all faculties starting October 2024. The aim is to visualize the similarity of reports and papers submitted by students on the learning management system (LMS) called manaba, comparing them with submissions from other students and internet information, thereby ensuring academic integrity.

• Osaka Medical and Pharmaceutical University: Has introduced "Turnitin Feedback Studio" at the university level and "iThenticate" for more advanced research. iThenticate is suitable for checking theses and papers by graduate students and researchers, as it primarily cross-references databases of academic journals and publishers.

As such, the initiatives of each university are diverse, regarding introduction timing, scale, and the adoption or non-adoption of AI detection features, indicating that there is no single "correct answer."

The biggest challenges: Japanese language support and AI detection accuracy

Turnitin's biggest challenge has long been "Japanese language support." Originally developed for English-speaking regions, the tool had limitations in terms of accuracy when dealing with the mix of kanji, hiragana, and katakana, the unique grammatical structure of Japanese, and the handling of academic terminology.

This situation took a major step forward in April 2025. Turnitin officially launched an AI writing detection feature that supports Japanese submissions . This new feature is trained to detect text generated by the latest language models such as GPT-4-o and GPT-4-o-mini, enabling more accurate analysis .

However, even with technological advancements, the fundamental challenge is not completely resolved. That is the risk of "false positives." Cases have been reported where texts written by humans with advanced writing skills are judged as AI-generated, or where the unique expressions of non-native speakers are incorrectly flagged. Turnitin itself acknowledges that the error rate increases when the AI score is in the low range of 1% to 20% . One study even shows shocking data that the false positive rate for texts written by non-English speakers reaches 61% .

For this reason, the mainstream approach in Japanese universities is not to treat the detection results as absolute evidence, but rather to use them as a "trigger for guidance." For example, the University of Tokyo has established guidelines to "not over-rely on detection tools" , and Keio University "refers to the detection results, but comprehensively judges after listening to the student's explanation." Cultivating academic integrity through human-to-human dialogue and education, rather than relying on technology, remains the most important approach even now.

University Responses — Guidelines, Training, and Data Privacy

To address the AI era, universities are not merely introducing tools but are implementing comprehensive measures.

• Guideline Formulation: Waseda University, the University of Tokyo, Keio University, and many other universities have published "Generative AI Use Guidelines." The common points among these guidelines are as follows:

  1. Do not uniformly prohibit: Acknowledge the utility of AI while ensuring a correct understanding of its risks (information leakage, copyright infringement, decline in critical thinking).

  2. Obligation to declare use: Mandate that any part where AI has been used must be clearly indicated, for example, in footnotes or references.

  3. Instructor's discretion: The final decision is entrusted to the discretion of the instructor in charge of the class. The permissibility of AI use must be clearly announced for each assignment.

  4. Training for Instructors: Simultaneously with the introduction of tools, training programs for instructors have become essential. Rikkyo University has published a manual for instructors, providing detailed guidance on Turnitin's operation, how to interpret AI detection results, and procedures for handling false positives. Kyoto University hosted a pedagogical practice forum in 2024 on the theme "What is better report creation support in the AI era?", creating a venue for instructors to share utilization methods among themselves.

• Consideration for Data Privacy: Turnitin is a U.S. company, and the fact that submitted student report data is stored on overseas servers has been a concern in relation to Japan's Act on the Protection of Personal Information (PIPA). For this reason, there is information that some universities, such as the University of Tokyo and Kyoto University, are adjusting their contracts to store data on servers within Japan. Furthermore, some universities are making considerations to leave students with options by making its use "optional" rather than "mandatory."

Academic Discussions and Future Prospects: Recommendations from the Science Council of Japan and Student Voices

Throughout the academic community, in-depth discussions are taking place regarding this issue. In February 2025, the Science Council of Japan published a recommendation entitled "Towards the Realization of a Society that Accepts and Utilizes Generative AI." The core of this recommendation lies in the recognition that "AI detection tools alone cannot guarantee academic integrity." Instead, it strongly emphasizes the importance of education that fosters students' ability to deepen their own thoughts and express them. This indicates a fundamental direction: not simply "cracking down" on misconduct, but rather building an educational environment that does not "give rise" to it in the first place.

On the other hand, what is the reality for students? Data from a 2024 survey indicates that approximately 50% of Japanese university students regularly use AI tools. At the same time, many students are strongly concerned about "disadvantages due to false positives." In fact, in 2023, an incident occurred at Waseda University in which a handwritten paper was judged to have a high score by AI detection, and the university subsequently worked to clarify the explanation process for students in response.

The lesson to be learned from such cases is that detection results should not be directly linked to evaluation, but rather positioned as a "starting line for dialogue." It is important for instructors to ask students, "How did you think about and write this part?" based on the detection results, and to carefully listen to their thought process. This process itself is a valuable educational opportunity to cultivate students' critical thinking and metacognitive abilities.

Overview: Towards the Coexistence of Humans and Technology

The current state of Turnitin in Japanese universities is not simply a matter of introducing a "monitoring tool." Rather, it is the totality of a multifaceted and complex endeavor to reconfirm the essence of education — "the meaning of learning," "the dignity of creating knowledge," and "respect for others' intellectual property" — within a rapidly changing technological environment, and to pass these values on to the next generation.

Turnitin, as a powerful technology, is certainly an effective tool that visualizes the risks of plagiarism and inappropriate AI use, thereby reducing the burden on faculty. However, whether this technology is effectively utilized to enhance educational quality ultimately depends on "humans." How faculty interpret the guidelines and communicate them to students. How students understand AI as a tool and leverage it for their own learning. And how universities design an educational environment that fosters academic integrity, rather than merely imposing regulations.

Technology continues to evolve. The accuracy of AI detection will likely improve even further in the future. However, "integrity" and "a spirit of inquiry," which are fundamental to academia, can never be replaced by technology. What Japanese universities should aim for is not to use Turnitin as a "shield" to prevent misconduct, but rather as a "mirror" to create a space for dialogue where students and faculty can learn and grow together. The future of education, where humans and technology coexist, will precisely be built upon this balance.