Every Day Seems Like Murder Here: Implications of a Rap Neural Network

Every Day Seems Like Murder Here: Implications of a Rap Neural Network makes use of three large databases to ask when and how one person’s lyric become another’s crime, and provide an interactive means of exploring the dynamics of the bias and stereotyping that makes this transformation possible.

We have found rap lyrics to be a highly productive foundation for exploring distinctions between art and commerce, bias from reasoned judgement, and the real from the fake. Implications of a Rap Neural Network provides the means to approach rap music as searchable and categorizable data, thereby disentangling its form, content, delivery and performance.

The first body of text is the Rap Almanac Database, a “big data” cultural research platform built around the cataloging, transcription and linguistic analysis of approximately 500,000 rap songs; Second is Rapbot, a machine learning algorithm trained on Rap Almanac Database content and a platform that generates rap lyrics on-demand based on user-provided keywords; Third is The State vs. Rap Lyrics, a digitized collection of court records in which the defendant’s rap lyrics were used as evidence against them.

The Project represents a unique opportunity to teach and pursue the kind of critical thinking and decision making that in a post-truth, Deepfake era, is a precursor to developing and deploying a healthy 21st century Democracy. We want to establish tools and methodologies for those who are most at risk of harm to have a say in how AI is used at the intersection of popular culture and the criminal justice system.