Digging for gold in the archive: Artificial intelligence is the treasure map

Abstract

There is gold (or at least golden nuggets) in digital archives, and unlocking that value can provide the new revenue stream that media outlets are looking for. Deployed wisely, artificial intelligence (AI) can offer the solution — it can scale quickly, it can have little to no impact on legacy workflow, and it can provide a direct line to monetisation. As this paper discusses, however, there are certain issues that must be addressed first, not least the fact that most media archives contain literally hundreds of thousands of video, image and audio files, for which metadata standards have evolved over decades. Compounding this is the lake of unused first-party data. This paper shows how AI can be used to solve the metadata issues underlying the various files in an archive and unlock the hidden value in so-called long-tail content, with minimal impact to existing work, workers and workflow.