AI Enables New Programming Strategies

AI’s continued integration into media asset management workflows enables organizations to reach into the past to fuel futurev stories.Not so long ago, the process of crafting a video package to mark the upcoming 60th anniversary of NASA’s first crewed space mission, Gemini 3, might have taken a content producer days of intensely manual work.

Dust off old archive tapes. Scrub through hours of historic news reports and vox pops to pull out compelling snippets and local angles. Isolate some NASA B-roll. Search for expert comment to add to the rough cut. Then, finally, it’s time to get oBut recent AI advancements eliminate the need for this old way of working. AI has reduced video sourcing workflows to a matter of minutes and made it possible for editorial teams to envision entirely new programming strategies.

Multimodal and generative AI have quickly become a must-have combination for indexing videos. This type of machine learning is designed to mimic human perception. It ingests and processes multiple data sources — including video, still images, speech, sound and text — to achieve a more detailed and nuanced understanding of media content, producing humanlike descriptions.

When integrated with a MAM, AI transforms the notoriously cumbersome search experience into something more akin to searching the web. When every shot in your media library is described in natural language, it unlocks the ability for you to talk to your media library in this same language. It’s made possible by an approach called a RAG (retrieval-augmented generation), connecting your full media library to a large language model (LLM). This enables producers to search for and find specific video shots and sequences in just seconds.

Latest AI developments involving the RAG approach go a step further and allow teams to surface the moments that are most relevant to their storytelling. For example, AI can now be customized to find and sort a season of a cooking show by top-rated dishes so that the footage can be swiftly worked into best-of episodes and snackable content.

Other new AI capabilities include the generation of custom video insights in the form of unique topic classifications and even article-length texts, headlines and summaries for OTT and social media. These eliminate manual video sorting and copywriting, which takes significant time.

We’re also seeing a strong adoption of AI voice cloning, and AI agents are quickly coming down the line. These will empower anyone in a media organization to request the video they want to work with — no need for specific search filters, simply an ask: “Show me Donald Trump’s latest White House press conference” or “Can you suggest today’s best Trump comments to turn into a YouTube short?”

Broadcasters and production houses are now using AI to mine their media libraries for new programming and revenue opportunities. One large media group is using AI to index 45 years of video archives, and pinpoint historic footage that can be reused as supplemental content to today’s news stories, to craft new documentaries and TV specials and to put up for sale and licensing.n with the creative part.