As AI chatbots become an everyday tool for information retrieval, researchers are drawing attention to the risks associated with using them to explain complex geopolitical issues. At the center of a new study is the situation surrounding Kosovo—a region with a long and contested history of international recognition. The authors examined how various chatbots formulate answers to questions about Kosovo's status, its history, and its relations with Serbia. It was found that artificial intelligence systems often simplify complex political realities, omit important historical context, or reproduce phrasing that can be perceived as supporting one side of the conflict. Researchers emphasize that such distortions do not arise from malicious intent on the part of developers, but rather from the characteristics of training models on large bodies of text, where different and sometimes contradictory viewpoints already exist. Since chatbots formulate answers in a confident and authoritative tone, users tend to perceive their words as objective truth without verifying sources. This problem becomes particularly significant as more and more people turn to AI systems instead of traditional search engines or news outlets. Whereas previously a user would see several links and could compare different viewpoints, they now receive a single, already synthesized answer, which increases the risk of one-sided perception of information. Experts urge AI system developers to implement mechanisms that allow them to point out the controversial or multifaceted nature of certain topics, especially those related to geopolitics, territorial conflicts, and historical disputes. The importance of media literacy among users is also emphasized—they should approach chatbot responses critically, especially on sensitive political matters.