The Persuasive Design of AI-synthesized Voices
(A2023-113984)

AUTHORS

Hannah Chang, Singapore Management University; Anirban Mukherjee, Cornell University

ABSTRACT

We investigate the impact of AI-based, machine-synthesized narrating voices on consumer cognitions and behavior in media-rich environment. Across four studies (plus pretests), we show that the design of AI voices systematically and predictably affects consumer cognition and behavior. Specifically, the designs of AI voices have differential effects in early versus later stages of consumer purchase journey. In situations where the consumers’ attention is already directed to the message, we find that marcomm with more AI voices generates a smaller proportion of favorable thoughts, which leads to a lower purchase likelihood. These results support our conceptualization that hearing more AI voices narrate a message is more cognitively effortful for listeners to process compared to hearing a single AI voice narrate the same message. Moreover, this effect is attenuated for consumers who enjoy expending cognitive effort and detrimental in consumption contexts where the consumer is more familiar with the product category. Substantive and theoretical implications are discussed.