<p>Artificial intelligence (AI) is increasingly promoted as a transformative force in smart city development, yet public narratives, particularly on platforms such as YouTube often blur the line between realistic innovation and techno-utopian idealism. This study critically examines the portrayal of AI-driven smart cities in the 57 most-viewed YouTube videos on the subject, using a mixed-methods approach that combines qualitative content analysis (via NVivo) and sentiment analysis (using the RoBERTa model). The analysis is structured around seven ‘smart reality checks’, namely transparency, financial viability, scalability, problem-solving orientation, regulatory compliance, social and cultural acceptance, and sustainability. The findings revealed that: (a) while technical aspects of AI are portrayed with a degree of realism, broader social and governance considerations are often neglected in favour of idealised narratives; (b) sentiment analysis indicates a complex public response; optimism remains strong, but there is a noticeable rise in neutral sentiment, signalling growing scepticism and epistemic ambivalence; (c) the rise in neutral sentiment in YouTube discourse indicates that audiences are becoming less blindly optimistic and more cautious or uncertain about the promises of AI-driven urban futures. By offering a critical evaluation framework for assessing public discourse on AI and smart cities, this study highlights the need to ground technological visions in context-sensitive, achievable outcomes rather than idealised projections, contributing to a more balanced understanding of AI’s role in urban transformation.</p>