About a month ago, rising film studio STX quietly put out a movie called Playmobil, inspired by the eponymous line of nonlethal vehicular toys for small children. Playmobil had a budget of $40 million and made back less than $14 million of it, worldwide. Before that, the company’s similarly toy-spawned Ugly Dolls lost many millions as well.
Although it probably shouldn’t take a predictive algorithm to ensure STX’s execs don’t make such financially grave miscalculations again, they’re not taking any chances. Last fall, STX became a client of Cinelytic, a company that offers AI for navigating decisions around prospective film projects. It’s a questionable move, but one that doesn’t seem out of line for a five-year-old film studio willing to try anything.
What’s much more troubling is this week’s announcement that Warner Bros. has just signed a deal with Cinelytic as well.
According to the Hollywood Reporter, the four-year-old Cinelytic has been beta testing its platform for three years and recently raised $2.25 million in funding. The system’s data-rich predictive analytics crunch the numbers on several factors, attempting to quantify intangibles such as star power, to assess whether a project seems worthy of green-lighting, and if so, whether it merits a theatrical release or is more of a streamer.
Although the article announcing the deal takes pains to stress that the AI is merely a time-saving tool to reduce executives’ calculative legwork, and is not meant to replace human brain-based creative decisions, this premise is fundamentally flawed. The kind of cut-and-dried financial decisions the AI makes are designed to have an impact on creative decisions. The software may not be Frankensteining together a patchwork quilt of random cinematic elements moviegoers like per se (while simultaneously eliminating the element of risk-taking), but that’s basically what we’ll be getting.
On paper, Warner Bros.’s prize director, Christopher Nolan, makes movies that sound too complicated to work. They only get green-lit and outfitted with enormous budgets because the analytics conclude that Christopher Nolan = Macy’s-parade-float-size dollar signs. But Christopher Nolan only became Christopher Nolan after a studio (Summit Films, a subsidiary of Lionsgate) green-lit his second movie, Memento. Would the algorithm have concluded that the backwards memory noir directed by some nobody starring no big-name actors was worth doing? If only Fast Company had an in-house AI so we could know for sure.
Risk-taking is part of the formula for cinematic success. Sometimes you lay an egg with the unconscionably plodding Goldfinch, as Warner Bros. did last fall, but if you don’t get fired for it, perhaps you also decide that a horror-adjacent take on the kid-friendly comic book movie, Shazam, should exist, as Warner did last spring, and live to take more risks later.
More major studios using predictive AI in their decision-making can only usher in the inevitable robot uprising, a scenario famously depicted in James Cameron’s original Terminator, a film any machine-based Magic 8-Ball would probably conclude was not economically viable.