Public opinion studies have observed that dissatisfaction with politics and the functioning of democracy is on the rise in numerous countries (Kriesi, 2020). While comprehensive diagnoses are necessarily multicausal, this phenomenon seems to correlate with growing percentages of the population that cast doubts about the ability of elected representatives and politicians, specially form mainstream parties, to respond to the demands of the electorate. However, the representative role exercised by conventional politicians can be challenged at least from two different perspectives, a populist (that focuses on the people) and a technocratic one (that focuses on experts), each one with its own set of antecedents and consequences (Bertsou and Caramani, 2020). This paper contributes to this discussion by measuring citizens’ attitudes towards a political system in which Artificial Intelligence (AI) tools are used as part of the institutional design. It also explores the extent to which attitudes towards AI-democracy associate with populist and technocratic attitudes. To do that we use an innovative measurement of attitudes towards AI-democracy that we incorporated in a panel study conducted in Chile, Spain, and the United States. We thus exploit variation between countries and across time to examine both the consistency of the main associations and speculate about the causal associations between key variables.