The advertising industry has been scared of AI, machine learning and cognition and its encroachment into creative territory. AI represents something intangible and plays into our deep-seeded sci-fi fears of machine overlords. But there is no doubt that in that fear is expansive opportunity to co-create and evolve how we create in partnership with AI.
To explore how we could use AI and machine learning, we set out to re-interpret how we could use technology to create music. Welcome TravisBott, an artificial neural network that we’ve creative-directed to make rap music.
Using the music library of Travis Scott, we influenced and trained our machine to produce original beats and lyrics. Our team taught the machine both grammar and rhyming cadence, building custom workflows to match rhyming couplets and number of syllables in each line.
We extracted the melodies, chords, and beats of music from Travis Scott to create a series of midi files, the digital building blocks of music. From there, we fed those files to TravisBott to begin creating music phrases, transposing the midi files into numeric sequences the machine could read and use to generate unique sequences. Using the minor key chord progressions and diatonic melodies Travis Scott’s music is known for, TravisBott was able to generate surprisingly similar music.