Understanding public narratives on contentious topics like vaccination adherence is vital for promoting cooperative behaviors. During the COVID-19 pandemic, significant polarization arose from concerns about vaccines, with misinformation and conspiracy beliefs proliferating on social media. While many studies have analyzed these narratives, the focus has largely been on English-language content. This linguistic bias limits comprehensive global insights. Our study introduces a novel multilingual approach that addresses this gap. By integrating Italian examples into a primarily English dataset, we detect vaccine-hesitant language and demonstrate the model’s adaptability to diverse linguistic data. Our findings highlight the importance of incorporating varied linguistic datasets for a more holistic understanding of global narratives on vaccine hesitancy.