Copy, claims and limits

Distribution copy turned out to be part of the artistic frame. Teasers were limited to 256 characters and descriptions to 1024; distributors also asked for sales facts covering PR, social plans, print, online media, touring, DJ feedback, airplay, direct-to-fan and advertising. Writing "the French version of…" or "part of a multilingual experiment" in any of those fields would have turned the release into derivative content at once. So the French act was presented as a French act and the Korean act as a Korean act, and marketing plans were written in the local language.

The aim was not to hide the construction from anyone; it was to test whether a music concept could stand culturally on its own without constant explanation. That is why we document it openly here.

A second boundary mattered more: ambitious copy must not invent success. "Radio outreach is planned for alternative and student stations" is a campaign plan. "Supported by national radio" is a factual claim. Sales facts were written around what was planned — PR outreach, curator pitching, short-form content, paid social, seasonal editorial pitching, specialist radio, December campaigns — and never around playlist placements, press coverage or airplay that had not happened.

Campaigns were local in voice but not national in reach: a Serbian song can reach the diaspora, German alt-rock is relevant in Austria and Switzerland, French extends across the francophone world, Brazilian Portuguese has its own large market. A market can be a language area, a subculture or a playlist category, not just a border.

Finally, scale multiplied risk. A prompt for "a male rock voice" can, in theory, produce a voice that happens to resemble an existing artist, without anyone intending or recognising it. Our response: be transparent about AI-assisted production, state that we do not aim to imitate existing voices, and offer an easy way to contact us if someone believes they hear a resemblance.