In today's age of social media and smartphones, portraits—such as selfies or pictures of friends and family—are very frequently produced, shared and viewed images. Despite their prevalence, the psychological factors that characterize a ‘good’ photo—one that people will generally like, keep, and think is especially aesthetically pleasing—are not well understood. Here, we studied how a subtle change in facial expression (smiling) in portraits determines their aesthetic image value (beyond a more positive appearance of the depicted person). We used AI-based image processing tools in a broad set of portrait photographs and generated neutral and slightly smiling versions of the same pictures. Consistent across two experiments, portraits with a subtle smile increased both spontaneous aesthetic preferences in a swiping task as well as improving more explicit aesthetic ratings after prolonged viewing. Participants distinguished between aspects associated with image beauty and the depicted person's attractiveness, resulting in specific interactions between variables related to participant traits, image content, and task. Our study confirms that a subtle—and in this case fully artificial—smile reliably increases the aesthetic quality of portraits, illustrating how current image processing methods can target psychologically important variables and thereby increase the aesthetic value of photographs.