Resources
Licensing research interviews and field recordings to AI
License interviews and field recordings the honest way: participant consent, ethics alignment, de-identification, ownership.
Updated 2026-10-27 · 5 min read
Research, interview and field recordings need a review of the original promises and intended new use before licensing is explored. Publication, recording consent and permission for a commercial training-data licence are different questions.
The review may involve participant terms, ethics approval, ownership, confidentiality and local requirements. Do not assume that de-identifying a transcript resolves every restriction on the original voice or video.
Read the source commitments
Identify the consent or release version, stated purpose, confidentiality promises and any ethics or sponsor conditions. Have the responsible institution or qualified reviewer assess the proposed use.
Record unresolved restrictions by session or participant group rather than treating the whole archive as uniformly cleared. Do not contact participants or promise revised uses without the appropriate authority.
Keep channels and representations explicit
Inventory original audio, video, isolated tracks, transcripts and field notes. Record language, recording conditions, session duration and edits. Separate researcher commentary from participant speech in labels.
An illustrative metadata record can include session_id, language, channel_map, transcript_method, consent_reference and restriction_status. Keep private identities and underlying permission documents out of a broadly shared manifest.
Plan privacy and controlled review
Consider identifiable voice, distinctive events and indirect identifiers in addition to names. A pseudonymous session ID does not by itself make a recording anonymous. Use the applicable review process for audio, text and linked metadata.
Share a non-sensitive description first. A sample should follow a controlled process and reflect actual limitations. The cited UK ICO guidance explains useful distinctions, but does not determine the lawful use of a particular research collection worldwide.
License only a reviewed scope
Define the exact subset, representations, permitted use and exclusions. Retain the source and transformation history so delivery can be reconciled with the approved scope.
Fiund’s standard buyer licences are non-exclusive and prohibit synthetic voice or likeness replication. An assessment does not establish that prior research permissions can be expanded, that a buyer exists or that payment will follow.
Frequently asked questions
Does this guide confirm a dataset is available?
No. Availability, permissions, included preparation and commercial terms require review of the actual proposed package.