A Robot’s Street Credibility: Modeling authenticity judgments for artificially generated Hip-Hop lyrics

TY - CHAP

T1 - A Robot’s Street Credibility: Modeling authenticity judgments for artificially generated Hip-Hop lyrics

AU - Manjavacas, Enrique

AU - Kestemont, Mike

AU - Karsdorp, F.B.

PY - 2019

Y1 - 2019

N2 - This study aims to advance and enhance our understanding of the properties that contribute to the perceived authenticity of a specific art form: Hip-Hop lyrics. The basis of our study is an experiment carried out in the context of a large, mainstream contemporary music festival. We crowdsourced a large dataset of authenticity judgements for both authentic and neurally generated Hip-Hop lyrics, which enable us to quantitatively assess human biases toward artificially generated text as well as which linguistic characteristics are perceived as authenticity cues. Additionally, the dataset provides solid ground for evaluating different neural language generation systems with respect to their perceived credibility. We compare a variety of language models and techniques. Our experiments contribute equally to improving the credibility of generated text and enhancing our understanding of the cognitive processes at play in the perception of authentic and artificial art.

AB - This study aims to advance and enhance our understanding of the properties that contribute to the perceived authenticity of a specific art form: Hip-Hop lyrics. The basis of our study is an experiment carried out in the context of a large, mainstream contemporary music festival. We crowdsourced a large dataset of authenticity judgements for both authentic and neurally generated Hip-Hop lyrics, which enable us to quantitatively assess human biases toward artificially generated text as well as which linguistic characteristics are perceived as authenticity cues. Additionally, the dataset provides solid ground for evaluating different neural language generation systems with respect to their perceived credibility. We compare a variety of language models and techniques. Our experiments contribute equally to improving the credibility of generated text and enhancing our understanding of the cognitive processes at play in the perception of authentic and artificial art.

M3 - Contribution to conference proceedings

BT - Proceedings of the 2019 Digital Humanities conference

ER -