This paper reports on a direct test of social responses to communication technologies theory (SRCT) with mobile messaging. SRCT predicts that people will mindlessly respond to computers in social ways that mirror their responses to humans. A field experiment (N=71) using participants' own mobile phones compared three influence strategies (direct request, flattery, and social norms) in the context of asking intimate questions of participants. These messages came from either an ostensibly human or computer sender. Flattery significantly increased self-disclosure when ostensibly sent by a human, but not when from a computer. The interaction effect for sender and influence strategy is inconsistent with SRCT's predictions. Implications for theories of source orientation, research methods, and future research are discussed.