This study by Min Ho An et al. aims to analyze public perception of antidepressants on social media using artificial intelligence. BERTopic was used to cluster discussion threads from Reddit containing antidepressant-related keywords, from a time period of 2009 to 2022. A RoBERTa model was then used to classify each discussion’s general sentiment, and the Mann-Kendall test was used to examine whether sentiment and topic proportion changed over time. The authors found that, while negative sentiment was prevalent, negative sentiment appeared to decrease over time, whereas both neutral and positive sentiment showed increase. Furthermore, topic clusters regarding the side effects of antidepressants saw declines in their relative proportion over time. Bupropion-related topics in specific showed an increase in discussion of therapeutic benefits and a decrease in seizure risk discussion. This gradual rise of neutral and positive sentiment may reflect shifts in public attitude. However, there is still a gap in understanding the longitudinal shift of these perceptions.