AI and hip-hop 

Hip-hop is a music genre that emerged in the USA in the 1980s. It combines music loops with rap and breakdancing. Later on, hip-hop fused with other musical styles and the content also changed: While early, or “old school”, hip-hop decried social grievances, a rap scene later evolved that was built instead on luxury and social prestige and favoured pleasing, pop-like beats. In China, hip-hop didn’t become popular until the 2000s, and it varies greatly from region to region.
AI can filter out similarities and differences in such a global musical style based on the variability of the dynamics, i.e., the extent and frequency of changes in volume. It can also detect the differences between individual pieces of music. But it cannot (yet) tell us the reasons for these differences.

AI sorts Hip-Hop

On the Kohonen map, you can see how the AI sorts hip-hop music. The dots on the graph indicate individual songs, with blue dots representing Western hip-hop and red dots representing Chinese songs.

Touch the dots to listen to the music or watch the music video!

Do you notice ‘outliers’, e.g., Chinese tracks surrounded by Western ones? Can you understand these ‘mistakes’ of the AI? It may help to operate the switches at the top left. Thus, you can observe the distribution of individual musical parameters on the map (where yellow indicates a strong influence).


Bader, R., Zielke, A. & Franke, J. (2021). Timbre-based machine learning of clustering Chinese
and Western Hip Hop music. AES Convention Paper (10473).