Brave New Worlds: Reimagining unscripted editing with AI
Published in Realscreen March 2024
New technologies such as artificial intelligence are seen by some as incredible tools that can take creativity to new heights, while others view them as developments to keep a careful eye on. As their uses accelerate within the content creation process, producers may find that both things can be true.
In this series, veteran unscripted media executive Fred Grinstein explores the brave new world of the rapidly evolving tech landscape, and how it will impact the unscripted production industry in the immediate and near future.
As any reality docusoap producer can verify, one of the challenges of making high-quality unscripted content is capturing the real-life drama of an ensemble cast with the fixed timelines of a production cycle. The steps in the process — shooting, editing and delivering to air — just don’t always keep up with the pace of real life, and particularly the frenetic speed of social media. Whether it was Kim Kardashian’s wedding that didn’t air on TV until a few months after the fact, or the recent kerfuffles surrounding “#Scandoval” on Vanderpump Rules, the optimal content experience for fans requires a time machine that doesn’t exist.
That said, quick-turnaround TV has been made before, and often the compromise is either a trade-off in quality or major costs that the network ends up having to weigh into its ROI. But what if there was a way to meet viewers where they want to be — or at least a lot closer to where they want to be — without breaking the bank?
Veteran unscripted producer Irad Eyal, co-founder of L.A.-based prodco Haymaker, is no stranger to docusoap drama or the pressures of delivering it to demanding networks. His past credits include spin-offs of Real Housewives of Atlanta, Secrets and Wives and Southern Charm for Bravo, as well as the global hit competition series Floor is Lava for Netflix. His latest pitch, however, isn’t for a TV show, but for an AI post-production tool called Quickture — a venture for which he serves as CEO.
“Our big mission is to transform the way that you edit so that you’re not using the paradigms of cutting film, [where] the icon is literally scissors for cutting an endpoint,” Eyal (pictured left) tells me. “We want to allow editors and producers to edit on a higher level, on the level of a beat or the level of a scene.”
Quickture works with Adobe Premiere Pro and Avid, both of which already have some AI-enabled features like transcript search. The software delivers new functionality, using their AI language models to first “analyze” the transcribed content through a variety of filters, assessing story value, emotional value, even a “spiciness” value. These color-coded and customizable criteria allow the creator to see at a glance what lies within their transcript beat by beat. And, much like AI note-taking tools such as Otter or Fathom, it offers the user a concise summary, the kind of top-line breakdown your field producer might be expected to send in after a day of shooting. All of this is done in a matter of minutes.
With this analysis completed, Quickture can then “auto-edit” based on your selected ideal running time. Let’s say you want a five-minute version of the hour-long interview you just shot. A few clicks and a few minutes later, and voila.
You can then refine the edit. Through a chatbot-like dialog box, you can prompt your edit, effectively giving notes in the form of a natural-sounding conversation, no coding required. A few minutes later, you get your revision plus a set of “notes on the notes” in which the app explains to you how exactly it responded to each of your refinement prompts — much like the in-line notes you might email back to your network exec who wanted some changes.
Having used this software myself in the process of making the above video, I can attest to its speed and ability to take many steps out of the traditional process. Several hours were reduced to minutes, resulting in a really fast and easy “radio cut.”
Eyal gives another practical example of how AI-aided editing can ease workflow via a scene from a pilot he recently produced (pre-Quickture). He and the network were at odds over the story direction: his team was drawn to the moments in their shoot when the featured chef/entrepreneur was brought to tears recalling the recent death of their grandmother, and how it had inspired the restaurant they wanted to open. The network, on the other hand, was more interested in focusing on the financial challenges surrounding the opening, and the drama behind an upcoming preview dinner.
The back-and-forth, looking at these different versions of the same scene, ended up dragging out over three or four days. If he had Quickture or a similar tool back then, Eyal believes that comparison could have played out in a matter of hours with AI-assembled rough cuts. And that’s not to mention any number of other story directions or hybrid combinations.
“A tool like this allows you to find your story, explore your story, and revise really quickly, which should make us all better storytellers,” he maintains.
Quickture is currently in public beta stage, with select production companies using the product. The roadmap ahead aims to build more sophistication, including global search options with which the tool can analyze thousands of hours of footage versus the individual sequences currently in play, as well as visual analysis through which the software can make narrative choices around shot selection (one shot or two, wide shot or tight, etc.). But in the meantime, the ability to simply speed up the assembly and rough-cut stage is its own innovation, and one that Eyal believes could impact the way we connect with our beloved housewives and docusoap characters.
“You shoot for three months, and now you’re editing for the rest of the year to get a season out,” Eyal describes the current unscripted production reality. “[The] cast is continuing their lives and big things happen. ‘Scandoval’ is a perfect example, when the cameras aren’t rolling and you have to scramble to get back in there. But what if we could change that paradigm? What if you could turn around an episode in a week or two in sync with the social media and the conversation that TV has been missing out on? Now, we’re an active part of it.”
Eyal sees this technology’s application going well beyond docusoaps. “We’re building custom models for formatted shows right now. If a client comes to us and they have 10 finished episodes of a series and they’re about to shoot another 20, we can train an AI model specifically on those finished episodes.” He also thinks independent documentaries, often under-budgeted, are an important focus ahead. What if the filmmaker could free up more resources for the more creative part of editing versus the slog of assembly?
Quickture joins the growing list of AI post-production tools coming to market imminently. Many are in beta, some have already been put into practice. While some creatives are thinking about mining the creative gems that often emerge with the introduction of new innovations, entrepreneurs such as Eyal are taking a bet on the picks and shovels that may become essential in this gold rush.
Still, it’s important to differentiate these tools from their generative AI counterparts, which are currently the centerpieces in debates about copyright. Tools such as Quickture focus on automation that works entirely from the original content the user feeds into it.
And, of course, it’s worth noting that the introduction of such tools inevitably raises concerns about the impact on the workforce.
“Fear about the industry and how AI is going to affect it is well-founded and real,” Eyal says. “I feel for everybody whose job is going to be disrupted by AI. I am one of those people.
“The budgets were going down before AI became a thing,” he continues. “Our jobs will be disrupted and changed in ways that we can’t predict. But reality TV has always been scrappy. We’ve always had to figure out the tools faster than anyone else, and we have the ability to adapt. We’re great storytellers.”