<![CDATA[The development of generative artificial intelligence has accelerated the use of deepfakes in political communication while raising complex questions concerning the boundaries between freedom of expression, information manipulation, and legal accountability. These challenges arise because the synthetic nature of digital content does not, by itself, establish legal unlawfulness. This study aims to develop a legal categorisation framework for political deepfakes to determine the boundaries of legal accountability based on the characteristics and context of their use. The study employs normative legal research with qualitative and comparative legal approaches. Legal materials are analysed through six dimensions, namely nature, context, intent, deception, dissemination, and harm, which are subsequently mapped into normative categories according to their relevance to applicable legal norms. The findings demonstrate that political deepfakes should not be assessed through a binary distinction between legal and illegal content. The analysis establishes four normative categories: legitimate political expression, political satire, informational manipulation, and legally accountable deepfake. These categories form a continuum of legal responsibility that enables a more proportionate assessment of the legal relevance of synthetic political content. The stronger the elements of intent, deception, dissemination, and harm, the greater the potential for legal accountability. This framework provides a conceptual basis for developing regulation and law enforcement approaches that safeguard information integrity and democratic processes without imposing disproportionate restrictions on political freedom of expression.]]>
