When artificial intelligence and video metadata collide: A learning curve

Abstract

This article describes the development of a media asset management system supported by artificial intelligence that serves not just as a metadata hub, but also as a content discovery platform. The project sought to rewrite the way video metadata are applied to both digital video and immersive content; create a mobile and cloud-based platform that would support the application of this new metadata standard, as well as the search and discovery of assets embedded with in-frame, time-based metadata; build a proprietary artificial intelligence bot to generate this new metadata standard; deploy it internally, and bring it to market for third-party licensing — all in less than 24 months. This article discusses the setbacks and surprises encountered on this ultimately successful journey.