We examined the causal relationship between playing violent video games and increases in aggressiveness by using implicit measures of aggressiveness, which have become important for accurately predicting impulsive behavioral tendencies. Ninety‐six adults were randomly assigned to play one of three versions of a computer game that differed only with regard to game content (violent, peaceful, or abstract game), or to work on a reading task. In the games the environmental context, mouse gestures, and physiological arousal—as indicated by heart rate and skin conductance—were kept constant. In the violent game soldiers had to be shot, in the peaceful game sunflowers had to be watered, and the abstract game simply required clicking colored triangles. Five minutes of play did not alter trait aggressiveness, yet an Implicit Association Test detected a change in implicit aggressive self‐concept. Playing a violent game produced a significant increase in implicit aggressive self‐concept relative to playing a peaceful game. The well‐controlled study closes a gap in the research on the causality of the link between violence exposure in computer games and aggressiveness with specific regard to implicit measures. We discuss the significance of importing recent social–cognitive theory into aggression research and stress the need for further development of aggression‐related implicit measures. Aggr. Behav. 36:1–13, 2010. © 2009 Wiley‐Liss, Inc.