6 Startups Using AI For Movies & Entertainment
When not investigating the latest possible investment scams or drilling down into the viability of equity crowdfunding, we’re pondering the really important stuff. Like why does the plot of just about every Star Wars movie revolve around a giant space station that needs to be destroyed? And how could the Empire afford to build these enormous weapons at the cost of what must be trillions of dollars without going bankrupt? A starship alone—if geeks and financial analysts out there don’t mind us crossing fictional sci-fi universes—could easily cost $500 billion, as we recently learned. It seems that it doesn’t take much brainpower to write a Hollywood script these days. So, it’s not surprising to learn that artificial intelligence has been recruited by the entertainment industry to do that and much more.
Many of you may be familiar with how an AI named Benjamin wrote a rather perplexing screenplay for a short sci-fi movie last year. It made little sense but was still better than anything starring Shia LaBeouf. And you may have actually read a story-not-about-Elon-Musk in Futurism.com about AI algorithms creating paintings on par with Picasso. One of our recent favorites involves Bot Dylan, a machine-learning composer of Irish folk music.
These are all cool and creative applications of AI in entertainment, but there aren’t any startups that we know of employing machine learning to churn our screenplays or musical compositions. Maybe it’s against the rules of the Screen Actors Guild. Regardless, we did track down startups using AI in entertainment-related ways, from offering cultural recommendations tailored to each customer to analyzing screenplays to find the next blockbuster.
6 Startups Using AI For Movies & Entertainment
Customized AI Entertainment
New York City-based Qloo is the Netflix of … well, everything that Netflix does plus music, dining, fashion, books, travel, sports, people, brands and consumer products. Billing itself as “the cultural AI”, Qloo has taken in $14 million since it was founded five years ago, including a $6.5 million Series B earlier this month. Investors include A-list stars like Sir Elton John and Leonardo DiCaprio. Qloo’s AI algorithms predict the tastes for any target audience and map relationships within and between all the different cultural categories above.
It starts with data, of course. For example, for the 18 million books in print, Qloo says its AI platform understands the setting, characters, genres, sub-genres, and more than 400 attributes down to the weight of the hardcover edition. For movies, its data cover 3.2 million movies, actors and directors. It then builds individual people profiles—120 million and counting—based on everything that person has loved and hated on the internet, along with all of the other unstructured data out there. It then starts to map those opinions, find connections (see above) and extract insights on what will most likely entertain you or end up in your shopping cart.
Meanwhile, back on the West Coast, LA-based IRIS.TV has raised $8.5 million during its five-year run to help companies to engage viewers longer with video content. IRIS does this through a tool called Campaign Manager, which the company says allows in-stream placement of branded, premium videos. It claims customers have seen an average increase of 70 percent in video views retention. How? As one IRIS.TV official explains in a company blog:
“The information we passively collect during consumer video viewing experiences informs our machine learning, and helps populate more relevant videos going forward. Data creates our intelligent video distribution models, helps clients monetize their video archives, and helps set that content up for increased discovery within the publisher’s own platform.”
Those clients include companies like CBS, Time, Sports Illustrated and Playboy. We wonder if Pornhub has approached them yet, though most customer engagements there won’t last more than five minutes anyway.
At the Movies
Remember when Gene Siskel and Roger Ebert were alive? Too soon? Well, studios no longer have to wait for a thumbs up or thumbs down to see if their latest Titanic will float or sink. Companies like ScriptBook, which took in $1.2 million in Seed funding last year, are using AI to predict how well a movie will do at the box office just based on reading the screenplay. ScriptBook’s AI platform binged on 4,000 scripts and 10,000 movies during the first six months, according to an article on Ubergizmo. The algorithm then came up with a system to predict financial success or failure.
In one case, ScriptBook’s technology (called Script2Screen) analyzed the script for the sci-fi movie Passengers, a film about two people on the longest road trip ever. It predicted a U.S. box office of $118 million versus $100 million in domestic ticket sales. The Antwerp startup’s AI can also predict critical success. For Passengers, it projected a rating of 7.3 for moviegoers against an IMDb rating of 7.1. In another case, it analyzed 62 films released by Hollywood studios between 2015 and 2016. It “greenlighted” the 30 films that turned a profit while turning down 22 scripts that would turn out to be a financial loss. They all starred Shia LaBeouf. At least that’s what we assume.
Predictive analytics for AI entertainment is something of an ensemble cast. There’s also an Israeli startup called VaultML. No word on funding, but the company caught the eye of Fortune, which explained that the company’s 4CAST platform can project ticket sales from a screenplay or a movie trailer. Fortune reported that Vault spent two years training its neural-network algorithm on “30 years worth of box office revenues, film budgets, audience demographics, and casting information”. The company CEO boasted to the news site that his AI entertainment platform gets close to the box office sales about 75 percent of the time. That assumption is that the machine will continue to learn … that Shia LaBeouf is a horrible, horrible actor. The company has another product called Alpha Vault that uses AI for business analytics targeting the entertainment industry.
Boston-based Pilot Movies is another young startup applying algorithms to predict box office gold. The algorithms compare a potential film project against a database filled with every bit of information on every film released since 1990, according to a story in the Boston Globe. A “bootstrapped dream” of three twenty-somethings, Pilot Movies says its forecasts achieve more than 80 percent accuracy once a trailer has been released. You can see how it did on a recent movie weekend.
The company is also using its AI and big data power to analyze other aspects of the entertainment industry like women and minority representation in the film industry, something that will always be dictated by what audiences will pay money to see.
Out of Context
Sounding like the Finnish goddess of the sea, Finnish startup Valossa has taken in a modest $650,000 in Seed money to develop an AI platform capable of detecting and identifying people, visual and audio context, spoken topics, named entities, general themes and explicit content in video. It creates that metadata for every second of video using computer vision, machine learning, and natural language processing. It’s like what some companies do with still images but perhaps even more impressive, if it works as advertised. Below you can see what Val.ai can do.
Studio Daily reported that some 120 companies had been beta testing Val before its release earlier this year. The company markets the technology as a way to solve the problem of content discoverability in video, but Studio Daily notes that one obvious application would be for media companies to use the metadata for contextual advertising.
While we’re still far from AI creating its own blockbuster movie or making Shia LaBeouf look like a competent actor, machine learning is finally breaking into the entertainment business. What we don’t see yet are VCs breaking the bank with blockbuster investments in AI entertainment. We wouldn’t be surprised if a few months from now we’re writing about another dozen AI entertainment startups, like we have recently with AI in recruiting. Things are changing that quickly when it comes to the artificial intelligence landscape.
Despite what the pundits say, FAANG stocks (Facebook, Apple, Amazon, Netflix, Google) don't give you real exposure to AI. Read about 5 stocks that give you true pure-play exposure to AI in our guide to investing in Computer Vision companies, freely available to Nanalyze subscribers.