We are witnessing that artificial intelligence and large language models are becoming an increasingly important tool in the political sphere of society, influencing election campaigns, political decision-making, and communication between the state and society. The aim of the work is to identify the risks of using artificial intelligence in political processes in Russia. The article comprehensively uses discourse analysis of scientifi c publications and academic research related to the risks of using large language models in politics, the comparative legal research method, and scenario forecasting, which allowed identifying gaps in legislative regulation and highlighting the need to improve legislation. The focus of this work is on the risks of using artifi cial intelligence in politics, including the bias and politicization of large language models, manipulation of public opinion and the spread of deepfakes, the use for cyberattacks on election infrastructure, and the management of electoral behavior. The main conclusion of the work is the assertion that there are risks of covert manipulation of public opinion and individual behavior management, which are benefi cial to the creators and owners of artificial intelligence. This can lead to the loss of national digital sovereignty and the delegitimization of the political regime. In the context of the militarization of AI, the Russian scientifi c community, lawmakers, and the public need to step up efforts to counter foreign interference through the use of artificial intelligence in political processes in Russia. The theoretical significance of the work lies in clarifying and expanding understanding of the potential risks associated with the use of artificial intelligence and large language models in the political sphere of society.