Technology Record - Issue 33: Summer 2024

106 BY ALEX SMITH Media companies may have access to huge libraries of historical content but have often struggled to make full use of these assets in the past. Microsoft’s Andy Beach explains how active archives are providing a solution Media organisations are constantly in the process of creating new content, especially as demand from viewers has continued to increase. But what happens to content in the days, months and years after it’s been made, distributed and viewed? Is the content no longer valuable, or can media organisations find new ways of leveraging it? These are the questions at the heart of the drive towards better asset management within the media industry. The large, traditional broadcasters and studios possess vast archives built up over many years which contain large amounts of content that can be reused to create further value. Yet their size also poses a significant challenge. Searching manually through such a collection is time consuming, meaning that simply finding a relevant piece of content for a particular project can prove to be a major stumbling block. “Archives today can be in a pretty wide range of states,” explains Andy Beach, chief technology officer for media and entertainment at Microsoft. “The most basic version would be an archive that is just cataloguing the assets. Things may be very disorganised. You might just have basic information, such as programme titles, rather than rich, time-based metadata. That makes is hard for an organisation to know what the archive contains.” Without knowing exactly what is inside its archive, it’s difficult for a media organisation to make meaningful use of the content it possesses. Many are therefore exploring an alternative solution made possible by artificial intelligence – an active archive. In an active archive, AI is used to analyse content and create rich pools of metadata to identify it. This allows users to quickly search through the archive for content relevant to their current task. AI is the main foundational element of such a model, says Beach. “In order to make an active archive, you truly need AI,” he says. “Without that, it’s just asset management. You need machines to have the ability to programmatically understand the content so you can then derive search capability from it and FEATURE Bringing life archives to