Charlotte de Witte releases Azadi on KNTXT with AI vocalist NAVA, proceeds going to FEMENA
Charlotte de Witte released the Azadi EP on her KNTXT label on 5 November 2026. Its three tracks, Azadi, Omid and Tales of the Desert, feature NAVA, an AI generated Persian vocalist created by artist Farbod Mehr to represent Iranian women barred from singing solo in public. All proceeds go to the women's rights organisation FEMENA.
Charlotte de Witte has released Azadi, a three track EP on her own KNTXT label that pairs her techno with NAVA, a virtual Persian vocalist whose voice is generated with AI. The EP came out digitally on 5 November, and all proceeds go to FEMENA, an organisation that supports women human rights defenders and feminist movements across Southwest Asia, North Africa and Asia.
NAVA was created by Farbod Mehr, a London based artist born in Iran, as a way to give a voice to Iranian women who are barred from singing solo in public. The three tracks are Azadi, which means freedom in Persian, Omid, which means hope, and Tales of the Desert. Each one is credited as Charlotte de Witte featuring NAVA, which puts the virtual singer on the release like any other guest artist.
The project landed in the middle of a loud debate. Coverage around ADE framed the EP as a test case for authorship and credit when an AI voice sits on a major techno record, and reactions online have been mixed. Some listeners see a creative way to make a point about censorship without putting a real singer at risk. Others are uneasy about one of the biggest names in techno normalising AI vocals while producers and vocalists are still asking how training data is licensed and paid for.
What makes Azadi different from most AI music headlines is that the tool is not hidden. The vocalist is named, the creator behind it is credited, and the purpose is stated up front. That is close to the consent, credit and compensation principles that industry bodies, including the Association for Electronic Music, have been pushing as the baseline for AI in music. Whether you like the record or not, it is a cleaner template than an anonymous AI topline dropped onto a streaming service.
For artists and managers the lesson is practical. If you use an AI voice or an AI tool on a release, decide how it is credited before the artwork is final, check what your distributor and the stores require you to disclose, and be ready to explain the choice publicly, because fans will ask. If you are a vocalist, this is a moment to get your own position on paper: what you will license, how you want to be credited, and what you would never allow. The conversation around Azadi shows that audiences now care about where a voice came from.
One more detail for collectors: some reports mention a 12 inch vinyl pressing later in the year, but it has not been confirmed, so treat it as a maybe for now.