A concept making the rounds in AI circles is something called “liquid content.” The shorthand describes the act of reshaping facts, ideas, and expressions across mediums. The most well-known example is a feature within Google’s NotebookLM: Once you’ve filled a folder with various kinds of data, it can whip up a podcast about that data, enlisting a couple of cheery AI-generated voices to give you an overview, analysis, or debate.
What do 1,000 journalists and PR pros know about AI that you don't? They took AI Quick Start, a 1-hour live class from The Media Copilot. 94% satisfaction. Find out how to work smarter with AI in just 60 minutes. Next class May 8. Get 20% off with the code AIPRO: https://mediacopilot.ai/ai-quick-start/
Push the idea to its limit and you arrive at a vision where any piece a media company produces can flow into every other format on demand. Making a podcast? With the right tools and prompting, in mere minutes, it can be reimagined as a series of clips, a feature article, or even an interactive presentation. And for a traditional news publisher, the same archive of articles can fuel videos that previously got benched as too costly to bother with.
This is no longer a thought experiment. I recently attended a couple of industry conferences—the NAB Show and Adobe Summit—and tools that can intelligently translate one format into another are showing up all over the floor. Just two examples: Amagi showed off an AI system that can scan a live newscast, understand the different stories covered, and create short-form videos for each one on the fly, populating a TikTok or Instagram feed almost as soon as the news is out. Stringr‘s Genna system, meanwhile, can take any news article and turn it into a video, pulling photos and licensed clips from repositories like Getty to assemble the footage.
Repurposing isn’t new, of course. But now that artificial intelligence can do most of the heavy lifting—interpreting the content, determining how it’s best expressed in a new form, and then pulling all the levers to do the actual work—the work moves faster, costs less, and scales in ways that weren’t possible before.
Automation doesn’t remove the hard part
If you sense a “but” coming, your instincts are right. AI can be a great catalyst in reimagining content, but it doesn’t solve every problem associated with pushing into new formats, and it tends to invent a few of its own. The opportunity is real, but treating liquid content like a magic growth engine is a mistake. It’s better understood as a new production layer that needs careful tending. As media companies turn to AI to expand their content footprint, here are a few realities worth keeping in mind.
1. Generative content produces diminishing returns. A small but important distinction first: there’s a difference between using AI to assemble content and using it to create content. It’s particularly relevant in visual media, where accuracy in the imagery matters greatly.
Setting aside the obvious ethical knots that come with generative video in news, there’s a quieter issue: Audiences don’t respond to it in the same way. Inception Media is a podcast company based on AI-generated scripts and synthetic voices. It does respectable numbers, but they’re far below what it might get from human-driven shows.
AI is a powerful accelerant, but listeners and viewers still reward authenticity. A publisher dipping a first toe into podcasting or short-form video with synthetic content may be disappointed by the audience numbers. The safer route is to stick with non-generative content and simply use AI to assemble existing footage and imagery. But that still requires you to either produce or acquire that material, which eats into the cost savings the pitch deck promised.
2. Good AI needs good data. For AI to understand and interpret content reliably, the surrounding data has to be accurate and thorough. That means things like tags, categorization, metadata, dates, and notes (e.g., exactly who appears in a video) should all be present and correct.
Most operations have improved on this front, but the older you go in the archive, the less reliable the metadata gets, especially across system migrations that scrambled or lost it entirely. Messy data is the rule rather than the exception in media, and it will keep many outlets from getting full value out of their back catalogs.
3. Humans still run the show. AI is a tool that gets better and more versatile every day, but it’s still far from perfect. It can hallucinate and misinterpret, and because it lacks experience with the real world, it sometimes makes mistakes humans never would (pointing out that volleyball is hard to play without a ball, for instance). Audiences have low tolerance for slop or poor quality.
The point is simple: AI can do plenty, but it still needs people. And not just to spot-check the output. Venturing into new platforms requires more strategic thinking than simply putting the content out there. To zero in on just one use case: AI can do competent translation, but launching into a new market is still a long, deliberate exercise in management and care.
The back catalog gets more valuable
All that said, if you crack the playbook, AI as a content-repurposing engine has serious upside.
1. Archives are a gold mine. Most outlets will reshare evergreen “hits” on social media, which can drive a decent amount of views. AI takes that idea further: not just resharing an article once, but extracting the best parts and turning each “nugget” into its own video, gallery, or social post. It can also reinvent the “this day in history” trick by spotting current trends and pulling forward older stories that map cleanly onto them.
2. A way into younger audiences. Many small and midsize outlets simply haven’t had enough content to really monetize on a platform like YouTube or Instagram Reels. Success is often a numbers game, demanding regular posting to even have a hope of showing up in someone’s feed. AI-assembled video won’t compete with MrBeast for attention, but it does give a brand a foothold with younger viewers, 63% of whom primarily get their news from these platforms.
3. Lean teams can do the work. Venturing into a new platform used to require weeks of study, hiring dedicated staff, and building out a strategy. Now AI can accelerate all of that—not just the remixing itself. As already mentioned, humans still need to manage the process and have the final judgment over whatever’s produced, but building a content-remixing department won’t be nearly as expensive as a pivot to video.
The harder question hangs over all of it: will the ROI justify the effort? As more media adopts remixing strategies and agentic systems, the inevitable result will be a large increase in supply of repurposed content—especially video. More supply tends to mean less demand per piece, which thins audiences further. In aggregate, the revenue lift from a remixing strategy may end up modest.
There’s a wrinkle, though. For niche publications with few competitors, there’s less of a danger of saturating their market, and making a move to a multimedia strategy—on a smaller budget—might improve audience growth and retention with readers who prefer formats like video and podcasts. Local and regional outlets fall into the same bucket.
The dream of a general-purpose content engine that can reliably spin out engaging stories in any format is getting less fictional by the day. But the engine is still just an engine. Building a successful strategy around it requires intention, careful curation, and a strong understanding of both the audience and the platform they’re on. Liquid content is a powerful idea. Pouring it well is still an art.
A version of this column appears in Fast Company.
