Given the fast-growing popularity of the ChatGPT tool and the heightened risk of students using it in writing their theses, experts from the Information Processing Centre (IPC) have decided to help all theses supervisors in the country- they have added a new function to the Uniform Anti-Plagiarism System (UAS). From February 2024, every higher education institution or doctoral school can now completely free of charge examine whether a given thesis was written using artificial intelligence. The new IPC solution will undoubtedly help in the even more effective fight against plagiarism at all levels of study.
Artificial intelligence writing theses for students?
Text automatic generation tools are becoming increasingly popular due to ease of use and the capabilities they offer to users. The use of “machines” for writing texts, especially as support in improving quality and content correctness, increases the demand for such programs. Text creation by artificial intelligence tools can boost users’ productivity and creativity. However, there are attempts to generate the entire text or a significant part of it and sign it with one’s name. Such actions provoke justified controversy.
Tools like ChatGPT are also popular among students, but there is no data yet on the scale of this phenomenon. Academic teachers have expressed the need for a tool that would allow them to verify whether a student has used artificial intelligence assistance in writing their thesis. This is why experts from the Information Processing Centre have expanded the Uniform Anti-Plagiarism System.
Dr. Jarosław Protasiewicz, the director of the Information Processing Centre says, “The Uniform Anti-Plagiarism System (UAS) is a modern tool that was created in response to the expectations of the academic community in Poland and organized the anti-plagiarism market in Poland. Previously, there were many programs, based on different algorithms and reference databases. With the UAS made available to supervisors in 2019, all engineering, bachelor’s, master’s and Ph.D. theses are analyzed using the same tool.” Dr. Protasiewicz adds, “Thanks to the UAS, we provided more comparability than in previous years between diploma thesis originality verification results across all centers in Poland. However, recently a new challenge arose with the growing popularity of automatic content generation tools. That’s why we have immediately started working to add a function to the UAS that will help supervisors identify if a particular diploma thesis was written using tools like the popular ChatGPT.”
It should be noted that the introduction of the Uniform Anti-Plagiarism System was a huge success. In 2016, only 30-40% of theses were analyzed for plagiarism. Since 2019 every thesis has to be checked. The system is constantly being developed by IPC experts who are working on its expansion and modernization. The addition of a module that helps detect fragments in the text written using artificial intelligence is undoubtedly an important achievement in the fight against plagiarism.
How does the UAS detect artificial intelligence application?
This is not the final word from the experts at IPC responsible for UAS. They continuously develop tools to detect text written by AI. Along with the development of new possibilities for applying artificial intelligence to various content and analysis, newer detection solutions will also emerge. IPC experts constantly monitor and perfect the system – introducing both ready-made solutions and tools developed at IPC. Information is currently being collected from UAS users across Poland about the new solution. Such research will help IPC experts add new functions to the UAS even more effectively to support supervisors’ work.
The detection method implemented into the UAS system is based on the hypothesis that the greater the regularity in the text, the higher the likelihood that it was produced by a language model. This assumption comes from the applied algorithm, which is able to generate regular and predictable text.
“The model used in the UAS has been trained on a large set of textual data. Its operation can be best explained by using the Perplexity measure. According to our assumptions, text generated by a random choice of words should be characterized by a high Perplexity value, while a very template and predictable text has a low Perplexity value. In the case of the latter, there is a justified suspicion that it was created using artificial intelligence,” says Małgorzata Wartacz, the UAS project manager at IPC responsible for the system development. “We are constantly developing our system and listening to feedback from the academic community. We are expanding the number of reference databases and ensuring greater internationalization of our tool. Of course, the UAS does not provide a hundred percent certainty that a text is plagiarized or written by artificial intelligence – it only estimates the probability. Therefore, the final decision is always up to the promoter. They verify the report received from the UAS and determine whether a particular diploma thesis is original,” adds the IPC manager.
It’s worth noting that currently no solution on the market is better able to answer; was this text written by artificial intelligence? Even the creators of the ChatGPT technology do not have a tool that could accurately assess whether a work was written by a bot or a human.
UAS – More than 400 institutions and 100,000 users
The role that the UAS plays in higher education is extremely significant. The IPC tool is available in every institution in Poland authorized to confer professional or academic degrees. Over 1.5 million theses have been analyzed using this system since the UAS was launched in January 2019. An average of 12 such works fall on one supervisor. Over 440 institutions have used the UAS since the system was launched, and it already has over 100,000 users. The IPC system provides real support for the work of every promoter in Poland. The UAS’s impact on the entire higher education system is best demonstrated by existing statistics. So far, over 320,000 works identified by the system as potential plagiarism have not been accepted by the supervisors. This means that the UAS has prevented over 320,000 copyright infringements.
It is important for students to understand the seriousness of plagiarism and respect the ethical rules in force at universities. Educational institutions, on the other hand, need to be consistent in enforcing their regulations to guarantee academic integrity. Plagiarism at the university is a serious violation of ethical rules and academic regulations. Its consequences can lead to serious repercussions for the student, including expulsion from studies and disqualification from defense. Plagiarism with the use of artificial intelligence is harder to detect, leading to loss of individuality and originality of diploma theses. It can contribute to the spread of false or unverified information and the drawing of scientific conclusions based on false data. The addition by IPC experts of a new function to the UAS, which allows stating that a given text was probably written using artificial intelligence is undoubtedly an important step in the fight against plagiarism.
