Artificial Intelligence Could One Day Determine Which Films Get Made

According to the founder of Artificial Intelligence (AI) outfit ScriptBook, Sony Pictures could have saved a fortune from 2015 to 2017 by using the company’s algorithms instead of human beings to reject or greenlight movies.

Disrupting Film Production and Distribution

In a presentation at the Karlovy Vary Intl. Film Festival, ScriptBook founder Nadira Azermai said that by analyzing screenplays, ScriptBook retroactively identified as box-office failures 22 out of the 32 Sony movies that lost money in that period, during which Sony released a total of 62 movies.

“If Sony had used our system they could have eliminated 22 movies that failed financially,” said Azermai.

Welcome to the brave new world of AI and machine learning as it applies to Hollywood.

Many see in ScriptBook and similar AI systems the potential to destroy a major part of the film production and distribution ecosystem, displacing script readers and saving much of the money studios spend on test screenings, focus groups, and market research.

At its most basic, ScriptBook, founded in 2015 and based in Antwerp, Belgium, has created a tool that analyzes the text of screenplays to produce financial forecasting, or as Azermai grandly puts it, “Our mission is to revolutionize the business of storytelling by using AI to help producers, distributors, sales agents, and financiers assess their risk.”

The cloud-based system is already in use. In 2016 ScriptBook raised USD 1.4 million in venture capital to accelerate development.

The system works like this: ScriptBook users upload a PDF file of a screenplay into the system. About five minutes later they receive a detailed analysis of the project that, among other things:

• Predicts the MPAA rating, analyzes its characters, detecting the protagonists and antagonists;
• Assesses the emotions of each character;
• Predicts the target audience, including gender and race;
• And, most importantly, makes box office predictions.

Providing Intelligent Movie Metrics

“When we show this to customers, the first question is: How is it even possible to give a script to a computer and somehow it can come up with all these outputs?” said Michiel Ruelens, data scientist at ScriptBook.

The answer, he says, is based on the fast-developing field of machine learning, whereby the software is first instructed by humans, then takes over the learning process and builds huge databases that can be mined at astonishing speed.

ScriptBook’s software has been “trained” on a large dataset of 6,500 existing scripts, said Ruelens.

Distributors will also find added value in ScriptBook, Ruelens added. “They take on a lot of risks when they buy rights to multiple territories. They now rely on subjective decision-making, reading the script and going with what their gut says. But we want to mitigate the risk by adding objective parameters that will tell them far more. Expertise means a lot but it’s important to back it with metrics.”

ScriptBook says that when its system indicates a script should be greenlit, it has an 84 percent success rate. This, said Azermai, is three times greater than the accuracy rate of humans.

“It’s also a validation tool,” she added. “It can validate the decisions you make.”

And in the ‘MeToo’ age, said Ruelens, ScriptBook can help improve gender parity in movies. The software can detect, for example, whether a film passes the test of including at least two female characters having a conversation that is not about men. Also, it can measure how many dialogs are between two men, how many between two women, and how many between males and females.

Naturally, the system isn’t perfect. When asked to retroactively assess the box office potential of “La La Land,” ScriptBook “predicted” the film would gross USD 59 million, whereas in fact it did more than USD 100 million in business (the difference could be attributed to the movie’s multiple Oscar nominations and wins, of course). But the system “greenlit” the film anyway based on its modest production budget.

Azermai acknowledged that many people think technology will eventually kill creativity, but that’s not the case, she said. “ScriptBook’s AI will just kick out movies that follow certain formulas. It’s very good at picking out artistic movies that do well financially.”

ScriptBook charges about USD 5,000 to ingest the script and generate a report on a single film, but offers discounts for companies that want to evaluate multiple projects.

© Copyright 2018 Variety Media, LLC, a subsidiary of Penske Business Media, LLC.