Automobile manufacturers are making information systems available in all vehicles. Most systems are screen based, but increasingly systems are either hybrids (screen/speech) or speech based. Speech systems in vehicles may have advantages over screen based in-vehicle systems; literature suggests that speech can be less distracting than screen-based interactions. But voices carry socio-economic cues including indicators of gender, age, personality, emotional state, ethniticity, education and social status, which in turn influence the perception and uptake of the message. This is true also if the voice is from a computer or other technical device. In previous research on social responses to media using voice technology, it has been shown that people tend to react to them as if they were real people. This suggests that these factors are also important to consider when designing in-vehicle speech systems. In this paper we first present the basic research paradigm, and then present three experimental studies of how variations in speech based information systems in cars influence both attitude towards the information presented and driving performance in a driving simulator.