Bangladeshi roads remain bumpy year-round-automatically detecting road damage seems promising, but are we truly ready to trust machines with our safety? In a survey of 120 respondents, 87% frequently encountered potholes, cracks, rutting, faded crosswalk markings, and uneven road surfaces. Most attributed these issues to inadequate maintenance and low-quality construction materials. To address these challenges, our research explored automated road-damage detection using transfer learning techniques. However, the lack of a dedicated Bangladeshi dataset compliant with the global Road Damage Dataset (RDD) standards [1] posed significant limitations. To bridge this gap, we initiated a local dataset collection and annotation effort following the global RDD annotation guidelines. Although technical limitations impacted initial results, insights from the survey emphasize the importance of integrating user-centric (HCI) aspects, such as automation trustworthiness, user-interface preferences, and practical usability considerations. Our findings underline that successful deployment of automated road damage detection systems in Bangladesh will require not only technological advancements but also substantial attention to human-centered design principles.