Crowdsourcing, the process of shifting individual tasks to a large group, may be useful for health communication, making it more people-centered. We aimed to evaluate whether a crowdsourced video is noninferior to a social marketing video in promoting condom use. Men who have sex with men (≥16 years old, had condomless sex within 3 months) were recruited and randomly assigned to watch 1 of the 2 videos in 2015. The crowdsourced video was developed through an open contest, and the social marketing video was designed by using social marketing principles. Participants completed a baseline survey and follow-up surveys at 3 weeks and 3 months postintervention. The outcome was compared with a noninferiority margin of +10%. Among the 1173 participants, 907 (77%) and 791 (67%) completed the 3-week and 3-month follow-ups. At 3 weeks, condomless sex was reported by 146 (33.6%) of 434 participants and 153 (32.3%) 473 participants in the crowdsourced and social marketing arms, respectively. The crowdsourced intervention achieved noninferiority (estimated difference, +1.3%; 95% confidence interval, -4.8% to 7.4%). At 3 months, 196 (52.1%) of 376 individuals and 206 (49.6%) of 415 individuals reported condomless sex in the crowdsourced and social-marketing arms (estimated difference: +2.5%, 95% confidence interval, -4.5 to 9.5%). The 2 arms also had similar human immunodeficiency virus testing rates and other condom-related secondary outcomes. Our study demonstrates that crowdsourced message is noninferior to a social marketing intervention in promoting condom use among Chinese men who have sex with men. Crowdsourcing contests could have a wider reach than other approaches and create more people-centered intervention tools for human immunodeficiency virus control.