What licence questions should a lab resolve before using a generated release?

**Does access to a sample establish permission to train on it?**

A buyer should ask explicitly. Sample inspection, model training and recurring evaluation are distinct uses, and the permitted scope should be documented rather than inferred from the ability to download a file.

**Which downstream uses need clarification?**

Ask about fine-tuning, continued pretraining, shared evaluation services and publication of record excerpts. Clarify how the licence addresses derived task variants, annotations and model artefacts. These are questions for the proposed agreement, not assumptions about standard MirrorFirm terms.

**What about collaborators and hosted infrastructure?**

Describe which entities and systems will receive the generated material. A lab may use contractors for annotation or infrastructure providers for training. The rights discussion should cover those intended transfers and any restrictions that affect them.

**What is a useful first step?**

Prepare a use map showing ingestion, storage, training, evaluation and external sharing. Include public reports and downloadable examples if planned. This gives the licensing discussion a concrete basis and helps the buyer identify permissions it needs before building a pipeline around a release.