Generativní umělá inteligence jako nástroj pro detekci dezinformací

This bachelor thesis deals with the issue of disinformation and artificial intelligence, more precisely with the possibilities of using generative artificial intelligence as a tool in fake news detection and information verification. Considering the current global political situation and the rapid development of artificial intelligence capabilities, this is an extremely topical issue. The theoretical part of the thesis is first devoted to disinformation, including an analysis of different definitions of related terms, historical and contemporary examples of its spread and basic strategies for its suppression. Subsequently, the principles of generative artificial intelligence, its architecture, insights into its future development and the current role of AI in detecting and spreading false information are explored, with an emphasis on its deployment in fact-checking processes. The thesis also includes the configuration of a custom generative AI model, whose goal is to provide a detailed evaluation of the truthfulness of articles and social media posts. To assess the accuracy of this tool, its outputs are then compared with the verdicts of certified fact-checking portals. Finally, a comparative analysis of currently available large language models was conducted using classification based on their accuracy in evaluating the veracity of statements made by political and other public figures.