The advent of computer-generated content offers various new possibilities in the marketplace. In recent years, billions of pieces of content have been automatically generated. However, the limited prior research on this topic mainly focuses on consumers perceptions in the field of robot-journalism rather than in potential marketing matters. To address this shortcoming, two experiments in a marketing context have been conducted to study the effects that arise when content is attributed either to a computer algorithm or to a human being. Results of both experiments reveal that trustworthiness, goodwill, and similarity mediate the effect of source on purchase intention, with consumers rating the human source better. Adding a byline explaining the functionality of an algorithm to the nonhuman source had no significant effects. Furthermore, several moderation and moderated mediation analyses provide additional insights and show that individual differences may be crucial. The present research thus extends understanding of the effects of computer-generated content and offers implications for marketers, who need to decide whether a computer algorithm or a human being is more suitable as a (declared) source.