For modern media organisations that store immense volumes of content, AI-driven metadata generation has become an essential capability. Replacing less comprehensive (and more costly) manual logging with automated transcription and video intelligence gathering, organisations can generate and organise a wealth of metadata that informs faster, more extensive searches for richer, deeper content repurposing. Metadata captured by various AI-driven analyses also facilitates rapid execution of key tasks in the area of compliance, be it meeting financial or contractual obligations around royalties for actor appearances or adhering to various markets’ differing standards for acceptable content (in audio and video).

HOW AI-GENERATED METADATA TRANSFORMS PRODUCTION

In the case of production workflows, AI can come into play right from the start. At a news network, for example, a speech-to-text engine can be used to analyse incoming feeds and automatically generate both a transcript and a brief description for each piece of content. With near-instant access to self-describing feeds, producers and editors can more quickly find the clips they need to tell a story, and gain time for the more creative aspects of bringing a compelling story to air.

During production of a reality TV show, dozens of cameras capture perhaps ten or 20 times the amount of content that will be used in the end product. Rather than task staff members with logging all of that video, the company can use AI-enabled tools to apply facial recognition and object recognition, as well as speech-to-text processing, to generate metadata automatically that effectively tells producers and editors who is in a scene, where they are, what they’re saying, and more. With this information in hand, the production team can jump into editing much more quickly and deliver a finished product ready for broadcast.

THE CASE FOR AI IN COMPLIANCE

The explosion in streaming services worldwide has made it much easier to export content for viewing in other countries, but most content requires preparation for global distribution. AI and machine learning (ML) can be applied to content to detect information that may require editing in order to comply with regulations in certain markets. Automated analysis of video using object recognition can significantly reduce the time needed to find and edit scenes with alcohol or smoking, for example, so that content aligns with regional preferences or regulations, and can be monetised more broadly, quickly, and cost-effectively.

Ongoing innovation in the realm of AI and ML continues to expand the analytical capabilities available to media organisations. Automated language translation is just one example of a technology that has improved immensely in recent years and is becoming readily available for incorporation into the latest tools for content creators and distributors. Digital Nirvana brings expertise in media to make the power of AI and ML available and easily applicable to broadcast and production workflows.

BUILDING BETTER CONTENT PRODUCTION PIPELINES

Market demand for original content these days is massive, and the rise of content consumption on social media and streaming video services has created a world in which creative endeavours are richly rewarded. But those endeavours are gated by various essential processes within the typical production and distribution workflow. AI and ML bring radical efficiencies to today’s content production pipelines. As broadcasters and streaming services automate or enhance operations with AI-generated metadata, they are positioned to address consumer demand with even more content, from a broader array of content creators, and to do so more quickly.

The media industry already benefits from highly capable general- purpose processing engines. Going forward, these engines will give way to specialised and customised engines dedicated to different types of content such as news, sports, drama, etc. To further improve accuracy, engines will grow even more specific, and address the scenarios unique to a narrower category, such as a particular sports league.

While AI and ML will evolve and grow ever more valuable in supporting content production, they only deliver on that promise when they are built into easy-to-use tools and solutions engineered to support media workflows. The exciting news for content producers is that they do have access to such products today, and to the expertise of trusted vendors in implementing AI and ML – and the metadata they yield – to the greatest advantage. ■