Experts state that artificial intelligence can be utilized to a certain extent in writing academic texts, and AI applications can be beneficial during the learning, research, and information gathering stages. Emphasizing that texts must be reviewed and edited to prevent plagiarism and misinformation, ensuring AI does not engage in actions contrary to academic ethics, Res. Asst. Şeyma Yektar noted that uncontrolled use of AI applications like ChatGPT in writing should be avoided.
Res. Asst. Şeyma Yektar from the Software Engineering Department, Faculty of Engineering and Natural Sciences at Üsküdar Üniversitesi, who conducts studies in fields such as machine learning and image processing, evaluated how artificial intelligence applications are used in writing academic texts.
Stating that artificial intelligence and machine learning technologies, which hold a significant place among developing technologies, have rapidly advanced in recent years and begun to be used in many areas, Yektar said, 'These technologies can create outputs similar to those produced by humans and can significantly reduce human energy and time. The concept of artificial intelligence generally encompasses a set of technologies and techniques related to the ability of computer systems to perform tasks requiring human intelligence.'
Artificial intelligence can be utilized to a certain extent in academic writing…
Stating that artificial intelligence is the simulation of human intelligence in machines programmed to think and act like humans, Yektar continued:
'One type of chatbot that uses artificial intelligence to understand and respond to user inputs in a natural, human-like manner is ChatGPT. Using AI tools like ChatGPT also has the potential to greatly enhance the writing of review articles for scientists by increasing efficiency and quality. Artificial intelligence can be utilized to a certain extent in writing academic texts.
AI applications can be beneficial during the learning, research, and information gathering stages. However, to avoid actions contrary to academic ethics, its limitations must be kept in mind, and the generated text should be reviewed and edited to prevent plagiarism and misinformation, and uncontrolled use of ChatGPT in writing should be avoided.'
It can be used by students for writing assignments and for cheating in the academic field
Explaining that ChatGPT, an artificial chatbot, generates text by providing answers after being asked the right questions, Yektar said, 'Through this working mechanism, artificial intelligence can be used by students for writing their assignments and for cheating in the academic field. AI applications might seem like a useful way to provide general information and a starting point for students. However, the output obtained from AI applications should not be evaluated as a professional article prepared by subject matter experts, and one must be aware that the information received is not entirely accurate enough to be quoted or cited. We must realize that we are not the authors of a written text prepared by artificial intelligence with minimal personal effort, and we must also be aware that this action will be considered plagiarism.'
Can text prepared by artificial intelligence be detected?
Şeyma Yektar also noted that academic researchers are looking for ways to detect whether a sequence of words was generated by a program like ChatGPT, and that Edward Tian, a student at Princeton, released a similar experimental tool called GPTZero earlier this year, targeting educators. She explained that this tool is an AI detector that checks whether a document was written by a large language model like ChatGPT, performing a probabilistic analysis of whether a text was generated by ChatGPT by detecting its 'randomness' and 'variance' features.
AI detection tool added to plagiarism software…
Res. Asst. Şeyma Yektar stated that an 'AI Detection Tool' has also been added to the Turnitin program, which is made available under TUBİTAK ULAKBİM EKUAL and allows checking student assignments for incorrect citations and potential plagiarism, adding that the currently available tools capable of detecting AI writing are 'GPTZero, Turnitin, and RoBERTa Base OpenAI Detector.'
Promises 95% success
Yektar also stated that the point to note is that these tools promise approximately 95% success, not 100%, and emphasized that by methods such as replacing words in the text with their synonyms, the text can be differentiated, allowing it to evade detection by these tools when it comes to identifying AI-written text.
Şeyma Yektar continued:
'Artificial intelligence in education offers great potential to enhance teaching and learning, but it also brings significant ethical challenges that need careful consideration and addressing. The ethical concerns arising from the development of artificial intelligence are also quite diverse and complex.
These encompass a broad framework, from large-scale issues like job losses due to automation, environmental degradation, and increased inequalities, to how AI might affect our privacy, our ability to judge what is what, and our personal relationships.'
What problems can artificial intelligence bring to the agenda in education?
Yektar stated that examples of problems that artificial intelligence can bring to the agenda in education include bias, discrimination, privacy, data protection, transparency, explainability, accountability, and responsibility, adding, 'Addressing the ethical issues of artificial intelligence in education is important for many reasons. The ethical issues of artificial intelligence in education can affect the fundamental human rights and dignity of students and teachers; the quality and effectiveness of educational processes and outcomes; and the participation and equality of educational opportunities and benefits. Addressing the ethical issues of artificial intelligence in education requires a multi-stakeholder approach involving collaboration among educators, policymakers, researchers, developers, and other parameters within the AI ecosystem.'
ChatGPT dangers and opportunities
Yektar also noted that the opportunities provided by ChatGPT and the dangers it may cause can be evaluated in both educational and cybersecurity contexts, and said the following:
An opportunity for visually impaired students!
'Firstly, if we are to evaluate its opportunities, it can be said to increase accessibility to education. For example, ChatGPT can provide verbal responses for visually impaired students. In addition, it allows not only visually impaired students but also students who have difficulty writing or using a keyboard to verbalize their questions. Furthermore, it can explain various concepts in a simpler way for students with learning disabilities.'
ChatGPT’s main danger is academic integrity
Yektar stated that perhaps ChatGPT's main danger is academic integrity, expressing that many researchers and educators believe using ChatGPT for academic work and assignment writing will only encourage cheating and plagiarism.
Yektar stated that although the information provided by ChatGPT may appear reasonable and well-written, its absolute accuracy cannot be guaranteed. She also noted that despite being trained on a vast amount of data, ChatGPT is not fully competent in all subjects and may not provide good answers to specific topics.
Şeyma Yektar concluded her remarks by saying, 'Of course, we should benefit from the advancements of technology. The important thing is to be able to define the limits of this situation. We must not engage in actions contrary to academic ethics, always investigate the accuracy of the information presented to us, avoid the uncontrolled use of ChatGPT in writing, and be aware of the dangers that ChatGPT may cause, such as security, privacy, and data leakage.'