Personally, I like post-apocalyptic movies and TV shows, as they offer a glimpse into alternative realities. Regardless of the director or whether the apocalypse is caused by technology or a virus, there is one common theme – art and cultural artefacts are preserved and seen as sacred remnants to inspire a better future. 

This is why I’ve always strived to be part of a company that has a long-term vision and sustainable operational strategy, especially, in an industry built on human talent, trust, and rights. Voice-Swap is

that

company. We are driven to be part of the music and creative culture. And we apply technology

for it

.

When we think about AI in music, it’s easy to get caught up in the novelty—synthetic voices, genre-bending remixes, and endless creative possibilities. But behind every AI-generated voice lies a deeper question: Who owns that voice? And more importantly, who gets credit, and who gets paid? Especially when we take into consideration music’s complex rights ecosystem. We believe AI voice models introduce an entirely new layer: the synthetic voice, which is a derivative of a real singer’s identity.

That’s why, at Voice-Swap, we treat every AI voice model as a

rights-bearing asset

, a new form of Intellectual Property. We don’t scrape audio from the internet or lift vocals from streaming platforms using stem separation. Instead, we collaborate directly with artists, estates, and rightsholders. We have clear usage rights under clear licensing agreements that spell out what can and can’t be done with the resulting model and outputs. Moreover, each AI voice is attributed to a real-world singer. 

There’s a widespread misconception that AI voice models are no different from traditional sample packs. But let’s be clear: unlike static samples, AI voice models can generate

infinite

derivatives. If used irresponsibly, that scale of replication can easily dilute the creative economy and devalue the original singer’s voice.

That’s why we not only promote clearly licensed relationships during the input (training data) stage but also advocate a

continuous revenue-sharing model

: every time an AI voice model is used, revenue is generated for the underlying voice owner. This new economy is not only direct and transparent but also

scalable

. In addition, it enables attribution, which in the music industry is currency.

Attribution affects everything from royalties to publishing splits and performance rights. Although today’s royalty systems weren’t designed to track and credit AI-generated vocals, at Voice-Swap, we view and treat AI voice technology as a legitimate participant rather than a “barbarian at the gate”. By doing so, we pave the way for integrating AI voices into existing industry frameworks, used by DSPs and PROs. After all, what’s the longevity of AI technology if it doesn’t offer commercial value-add for both its users and contributors? 

When discussing rights and attribution, voice-to-voice conversion raises intriguing questions about ownership. Unlike text-to-speech, where the voice and the performance are solely based on the synthetic voice model, voice-to-voice conversion involves (at a minimum) two creative inputs: the melody and emotion from the user, and the vocal identity of the AI model. Both contribute something essential. Both deserve recognition. Both parties’ interests should be considered.

These are just some of the fundamental considerations being discussed internally at Voice-Swap to understand what to develop and how to apply the technology so that something perceived as counter-culture can become pro-culture. Guardrails against misuse of one's voice, licencing, attribution, and compensation are responsibilities

we

, as tech developers, must consider, as AI voice technology unveils a plethora of both creative and practical possibilities across various domains.  

Moreover, when all is said and done (referencing back to my soft spot for post-apocalyptic scenarios), we want to look back at what we built and be remembered not as part of the noise or contributors to AI fatigue, but as part of the music culture, built on shared values.