FussAI is opening its Contributor Program to authors, bands, singers, producers and instrumentalists. Material is not quietly swept into training. Rights come first, then an electronic agreement, review, and only then possible dataset inclusion.
From positive Distributable Net Result, 10% goes directly to eligible Contributors by Contribution Units and another 5% is separately earmarked for OSA and INTERGRAM at 2.5% each. The collective allocation does not reduce the direct pool.
You can end future use of source files. We cannot promise that an already-trained contribution can be “unlearned” with one click. If that historical contribution remains in a commercially used model lineage, its Legacy Units continue to participate in the annual pool.
An individual signs for themselves. An authorised representative may sign for a legal entity. For an informal band, all listed members/rightsholders must confirm before FussAI can activate the agreement.
The current program is intended for human-created music and performances. AI-assisted elements must be disclosed and reviewed; fully AI-generated material is not accepted into this training program.
A general music model is not the same as a named model of a specific voice. Targeted imitation of a recognisable voice requires a separate explicit electronic addendum.
For every accepted contribution we aim to record the agreement, source, rights, model lineage, Contribution Units and change history.
The electronic signature creates an immutable snapshot of both CZ and EN wording. Depending on band structure, additional member signatures are required. Material then undergoes a separate rights and technical review. Nothing enters training without manual approval.
Non-binding contact only. This form does not grant a training licence or permission.
Sign in with Lajkto for a traceable request.
Sign in