Schopnost Gen Z rozlišovat mezi lidským a AI vytvořeným textem v závislosti na úrovni využívání generativní umělé inteligence

This bachelor's thesis deals with the ability of Generation Z to distinguish between human-written texts and texts generated by artificial intelligence tools in the Czech language. The main objective of the research was to determine whether the frequency of using generative AI tools and the level of user experience affect the success rate of detecting synthetic content. The theoretical part defines the functioning of large language models and the concept of AI literacy in the context of the target group. The practical part is based on a quantitative questionnaire survey in which respondents were tasked with identifying the authorship of text samples from the fields of news reporting, poetry, and internet forums. The research results showed that the overall success rate in detecting AI texts is only slightly above the level of chance (55%). Surprisingly, the findings revealed that neither a higher frequency of use nor advanced experience with AI tools leads to better detection results. Furthermore, the assumption that users would be more capable of identifying outputs from their preferred language model was not confirmed. The thesis concludes that contemporary AI-generated texts in Czech are practically indistinguishable for Generation Z, which highlights the necessity of active education and the continuous development of AI literacy within the educational system.