Dehumanization continues to be prevalent today and predicts detrimental intergroup consequences. Thus, it is important to identify novel interventions that reduce dehumanization and explore the mechanism(s)—both established (e.g., empathy induction, intergroup contact) and relatively understudied (e.g., humor)—driving the effects. To address this issue, in Study 1 (N = 2,349), we conducted an “intervention tournament” and found that a video (i.e., “Mean Tweets”) of a relatable and diverse group of young Muslims ridiculing Islamophobic comments posted on an online video of a Muslim preschool burning down effectively reduces hostility towards Muslims. Specifically, the Mean Tweets intervention significantly reduced dehumanization of Muslims and, although the effects were weaker, anti‐Muslim policy support. However, a follow‐up study conducted 1 month later revealed that these effects subsided. Next, in a preregistered study (i.e., Study 2) (N = 677), we find that our intervention reliably reduced dehumanization of Muslims, but the reduction of anti‐Muslim policy support was not replicated. While investigating our intervention's effect on anti‐Muslim policy support, we find that our intervention led participants to discount the intervention's intended message, which could have short‐circuited the intervention's effectiveness on policy support. Considering these results, we discuss the potential psychological processes (e.g., humor, message discounting, tone) underlying our dehumanization‐reducing intervention.