
REAL ML
REAL ML supports and advances interdisciplinary research that holds the use of algorithms and AI technologies to account—with a focus on the perspectives and experiences of marginalized communities
A campaign to create hack algorithmic music discovery algorithms to promote an anti-Nazi message and pressure audio platforms to stop hosting Nazi content
Neo-Nazis are more present in Germany than they have been for a long time. Music has always been the entry point into the scene for many. „Most of them were somehow recruited through music.” Philip Schlaffer, Ex-Nazi Unfortunately, today everyone can get Nazi songs as easily as never before: on music streaming platforms like Spotify, SoundCloud and Co. The situation is even worse: the platforms’ algorithms actively spread these songs and bands. Those who listen to Nazi music once get recommended more and more – and the platforms do nothing. We turn the tables and let the algorithms work against the extreme right. To do this, we developed a Trojan Nazi band: HetzJaeger. So perfectly tuned to the primary parameters of algorithms that it is automatically played to all listeners of far-right music. We started with a teaser, trained the algorithms with it for a month and released the complete song on the 89th anniversary of Hitler’s seizure of power.


REAL ML supports and advances interdisciplinary research that holds the use of algorithms and AI technologies to account—with a focus on the perspectives and experiences of marginalized communities

What if every AI analysis of public data came with a verifiable evidence record?

The Taurus is a platform for creating and managing transparency notices for political advertisements as required by EU Regulation 2024/900.

Academic research project on algorithmic risk assessments mediating access to private renting, social housing, mortgages.