Academic Integrity in the Age of Artificial Intelligence and the Validity of Violation Detection Tools

Abstract

Purpose of the study – to analyze the transformation of academic integrity principles under the conditions of rapid development of artificial intelligence (AI) and to determine the effectiveness and validity of modern tools for detecting academic misconduct in the digital educational environment. The article focuses on the challenges associated with the use of generative technologies in research and learning activities, as well as on the search for ways to adapt ethical and legal norms to the new conditions of academic communication. The research methods are based on the application of systemic, comparative, and analytical approaches, as well as content analysis of modern publications. The results of the study demonstrate that the use of AI simultaneously increases the productivity of academic work and creates risks of violating the principles of authorship, independence, and reliability of results. It was found that most existing detection tools show limited validity in identifying AI-generated texts, especially in the humanities, where contextual interpretation plays a significant role. A comparative analysis was conducted of detection algorithms used in Turnitin AI Detector, GPTZero, Copyleaks AI, and Writer.com, taking into account their principles of semantic text evaluation, training data sources of language models, and the accuracy levels in identifying AI-generated content. The conclusions justify the need to reconsider approaches to academic integrity as a dynamic system of ethical norms that must take technological change into account. The article proposes a concept for integrating AI into academic culture through the development of digital ethics,new criteria for assessing the independence of student and research work, and improved mechanisms for validating research results. It emphasizes the importance of developing national standards for the use of AI in education and promoting digital literacy among scholars and students. The obtained results can be used to modernize academic integrity policies, improve systems for checking academic works, and strengthen trust in research outcomes in the age of artificial intelligence.