Generative Artificial Intelligence and the Academic Integrity of Graduation Works in Economics - Exploring Perceptions of Romanian Academia

Technologies based on generative artificial intelligence (GAI) and large language models (LLM), available on platforms such as ChatGPT, have raised intense concerns about their impact on education. In the context of knowledge assessment, a particular challenge is to ensure academic integrity for bachelor's and master's theses, henceforth, graduation works (GWs). This article presents the results of an exploratory study on the perceptions of academia in Romanian economic higher education regarding the academic integrity of GWs in the context of using artificial intelligence (AI) tools based on LLMs, such as ChatGPT. The survey-based research provides a detailed snapshot of the target group's attitudes and perceptions; the study also identifies significant relationships between constructs and proposes two latent factors - a structural and a consultative dimension - that reflect two facets of LLM use in authoring GWs.