Rights subsistence: what AI companies need to know
Rights subsistence is one of the most important and least understood concepts in the AI-and-IP debate. Understanding it is not optional for AI companies. It is the foundation of every legal question about training data, output rights, and liability.
What subsistence means
In intellectual property law, a right "subsists" when it exists in a work automatically — without registration, without assertion, and without notice. Copyright in the UK and EU subsists from the moment of creation.
What AI does to subsisting rights
- Training data ingestion engages copyright, database rights, moral rights, NILP
- Fine-tuning engages copyright, neighbouring rights, NILP, biometric rights
- Generated outputs engage derivative rights, moral rights, NILP
- Distribution engages distribution rights, performance rights, attribution
The liability landscape
AI companies face subsisting rights liability at each pipeline stage. The CIP framework provides a structured path to managing that liability: through licensed ingestion, provenance tracking, and recognised compliance via Platform Certification.
What responsible AI companies are doing
- Licensing training data through recognised frameworks
- Maintaining Rights Payload records for ingested content
- Honouring TDM opt-outs declared via cip.md and robots.txt
- Seeking Platform Certification to demonstrate compliance