Lingua Libre lacks a Linked Data dimension and the Commons category as well as the EXIF data do not allow for a proper utilization of the metadata.
Each recording needs to be identified by an URI (the URI should not be the same as the one where the actual recording is on Commons);
We would need to have a simple ontology to describe basic terms for concepts related to languages: language, time of recording, (anonymized) locutor, background of the (anonymized) locutor, word (as written?), links to Wikimedia Commons (for the actual audio file), links to Wiktionary (for the word)...
The uploader and other contributors of Lingua Libre need to be able to describe their recordings, or batches of recordings, using the vocabulary (see two points above), and linking to other data either on Lingua Libre, and also on Wikidata.
Finally, the last step would be to publish our ontologies and those datasets, for example using simple download links that anyone can use.
Lingua Libre lacks a Linked Data dimension and the Commons category as well as the EXIF data do not allow for a proper utilization of the metadata.