Sourcemark makes AI training choices visible before training.

Creative work is moving through platforms, suppliers, crawlers and datasets at a scale that existing permission and licensing systems were not built to handle.

For creators and catalogues, there is still no consistent way to declare whether work is available by licence or unavailable for AI training in a form AI teams can check. For AI teams, there is no standard place to see whether a creator, catalogue or representative has declared a position before training or fine-tuning.

The result is avoidable uncertainty. Creator choice is often invisible, licence routes are hard to find and questions about permission surface too late, after work may already be inside a training workflow.

Sourcemark moves the permission check earlier, before creative work becomes a training data decision. It gives creators and catalogues a structured way to put AI training choices and licence routes on record, and gives AI teams a clearer way to check, route and document before training or fine-tuning.

Sourcemark sits between the people and organisations publishing or representing creative work and the AI teams checking candidate data before training or fine-tuning.

Creators and catalogues register works with AI training choices, fingerprints and licence routes. AI teams can check candidate datasets against Sourcemark and receive a result they can act on: available by licence, unavailable for AI training, hold for review or no registered signal.

Where Content Credentials or other provenance data are present, Sourcemark can work alongside them. Provenance helps show where a file came from or how it was edited. Sourcemark records the AI training choice and licence route so they can be checked later.