The emergence of artificial intelligence and the prevalence of recommendation algorithms in media platforms have challenged traditional paradigms of political participation. Using a descriptive-analytical method and a critical approach, this article seeks to explain the mediating and structural role of artificial intelligence algorithms in changing the nature of citizens’ political activism. The main question is how the technical architecture of social media, by prioritizing personalized data and the logic of the attention economy, directs political participation from the level of conscious and rational action towards reactive, emotional, and engineered actions. Algorithms, as hidden political actors, reorganize the perceptual structure of citizens and erode the foundations of communicative rationality and public space. The results show that political participation in the algorithmic era has suffered from a quantitative-qualitative paradox, namely a significant increase in interactions, but a profound decrease in the quality, representation, and legitimacy of political action. Therefore, the accountability of platforms towards algorithms influencing public opinion should be institutionalized as a fundamental principle in democratic digital governance in order to reclaim citizens’ political agency from computational power.