This study investigates public perceptions of Artificial Intelligence (AI) in Taiwan, focusing on the roles of political ideology, science news consumption, and knowledge. Analyzing responses from 502 participants, the research reveals that political ideology has a limited impact on AI benefit perceptions but likely becomes more significant with increased AI application. The study suggests that, for narrow AI, science news consumption and content knowledge are more influential than political ideology. It emphasizes the need for tailored communication strategies and highlights the positive association between respect for science authority and favorable AI perceptions. The findings provide essential insights for policymakers navigating AI adoption.