But, What If We Could Know Everything? An Analysis of The Emergence of Artificial Intelligence in Hollywood

By Zoe Ganek

Oscar-winning screenwriter William Goldman opened his 1982 memoir Adventures in the Screen Trade with four simple words: “No one knows anything.” This statement is regarded as one of the best lines in Hollywood history, as Goldman simply yet effectively summarized the unpredictable nature of the movie making business (Debruge). It makes sense why Goldman stated these words; many of his most famous scripts such as The Princess Bride (1987), Butch Cassidy and the Sundance Kid (1969), and All The President's Men (1976) were not expected to reach the level of success that they did while other films from the time period that were expected to be hits, flopped. Goldman’s specific and unconventional writing style was risky, as he was not afraid to break the rules of traditional screenwriting (as seen in the two act structure of All the Presidents Men and the whimsical elements of The Princess Bride.). Yet, the studios that produced his movies took a chance on Goldman, and, as a result, produced timeless and iconic films that are still relevant and enjoyed today (Debruge). While the unpredictability and high risk of the movie business is likely not seen as a positive for studio executives, it is for creative visionaries and audiences alike. Part of what makes films exciting is the possibility of the unexpected, and the unexpected success of an non-mainstream film (take Parasite (2019) or Everything Everywhere All At Once (2022) for example) further adds to the allure of Hollywood itself. Yet, for the people financially invested in film, this unpredictability is not as enjoyable when millions of dollars are on the line, and studios have been vying for a way to mitigate risk in the film industry throughout history. In a post-Covid, peak streaming era, the desire for box office blowouts is increasingly high. The emergence of A.I. (Artificial Intelligence) technologies offers studios a possible, albeit concerning, solution to the historic issue of “not knowing” anything and dwindling box office numbers. Through offering studios technology that presumes to analyze and predict the success or failure of a script or a film based on the input of past data, AI technology threatens to turn Hollywood into even more of a repetitive, formulaic echochamber than it has already become, perpetuating the industry’s issue of prioritizing profit over originality and creativity for the worse. 

In order to understand the potential market for AI technology, we must first look at the current state and dynamics of the film industry. Author Ben Fritz in his book The Big Picture; The Fight for the Future of Movie (2018) argues that in response to the consistent issue of “not knowing anything”, executives finally do know something, and it is that “branded franchises work” (Fritz, 187). Based on the list of the highest grossing films of the 2010s, Fritz’s assertion reigns true, as every single one is a franchise film, with 8 out of the 10 films being Disney films. Disney has undoubtedly been at the forefront of this franchise era of Hollywood, as the studio has made a point to focus exclusively on churning out branded franchise hits. Whether these types of films are more like “theme parks'' — as Scorsese called them — or pieces of influential cinema is up to debate, but it is indisputable that these films are the most widely consumed by global audiences, which in turn, means they are making the most money (Bell). Disney’s power does not only lie over the box office, but also the trends of mainstream Hollywood, as Fritz describes Disney as “the studio the rest of Hollywood is striving to emulate” due to the practically guaranteed success they have found with their business model of releasing a smaller slate of high budget franchise films (Fritz, 185). 

In 2018, the time that Fritz released his book, he was clearly lacking an optimism for the future of smaller budget original films at studios. He argues that the increase in “original risk taking television” is “directly tied to the decline of original, risk-taking filmmaking and the dawn of the franchise age of film” (Fritz, 200). The content produced and released by or on streaming platforms have an inherent safety blanket, as their success is not dependent on box office numbers or critical reviews. The Covid-19 pandemic, in its complete disruption of the movie theater business, cemented and expedited the power of streaming platforms and allowed for more risky films to be made. But while streamers were able to find success in their production and distribution of more artistically driven films, especially at the 2022 Oscars when Apple TV’s Coda (2021) won Best Picture and other streaming films were nominated, this was a unique exception, as the Oscars have now re-declared that theatrical releases are again required for a film to be considered by the academy. Furthermore, without theatrical releases streaming services will be more likely to have to continue increasing their prices as their beginning subscription fees are not economically sustainable in the long run (Dietrich). IMAX CEO Richard Genfold this past November made his opinion on this clear, asserting that the “the argument that you can skip a theatrical window and make up for it on streaming just isn't true” because “the theatrical window is more essential than ever to get the streaming revenue” (Genfold). While Genfold is likely to have some bias on this topic, his statement is suggestive of a near future in which streamers start having to implement theatrical releases. Since the types of films that dominate the box office are high-budget franchise films, it could also be suggestive of an uncertain future for films like Coda. From the unprecedented negative impacts that the pandemic had on the film industry, to the impenetrable success of Disney and franchise films at the box office, to the pressure being placed on streamers to have theatrical releases, I believe it is safe to say that studios and production companies are desperate for low-risk, high-reward films. AI technology offers a solution to this issue.

AI can have multiple applications in the pre- and post-production stages of filmmaking. Some softwares can assist in the editing process by using analysis of scripts and footage to splice clips together, while others can be used for storyboarding by feeding softwares specific scene descriptions (Raevskiy). Other softwares have gained attention for their ability to generate their own scripts through analyzing past narratives. The script of Sunspring, a sci-fi short made by Oscar Sharp for the Sci-Fi London 48 hour Film Competition, is an example of this as it was written completely by A.I (Raevskiy). Sharp fed a slew of sci-fi scripts into a predictive text network, and then instructed the software to create its own. While the idea that A.I could replace real human screenwriters is undoubtedly an unsettling idea, the film’s nonsensical and unnatural dialogue is demonstrative of not only how this technology is still in the very early stages, but also how this use of AI in filmmaking is more for spectacle and experimentation, rather than long-term use.

When one hears ‘AI’ and ‘Hollywood’ being used in the same sentence, images of scripts and films being magically made by robots might come to mind. However, while such softwares might exist, the real use of AI in the film industry seems to be occurring behind the scenes, as AI software has emerged as a tool for predicting a film's box office outcome. Companies such as Cineyltic and Scriptbook offer to assist filmmakers and studios in their decision making processes through utilizing analytical AI technology that can predict the projected success of a script or a film, recommend a target audience, and suggest which actors and directors would be the most economically beneficial choice. In January 2020, Warner Brothers publicly announced its partnership with Cineyltic with plans of employing its AI software to “inform decision making around content and talent valuation to support release strategies” (Rose). This was the first time a major studio like Warner Brothers had announced its use of AI, but Sony Pictures’ Ingenious Media, and STX entertainment are also clients of this company. Cineyltic claims to have gathered data from over 95,000 movies and 500,000 actors and to have a 85% success rate at predicting box office outcomes. The company’s biggest “selling point” is that its system enables clients to easily see how a predicted success rate changes based on the chosen actor, genre or release strategy (Rose). Journalist Steve Rose describes Cinetylic’s AI as “fantasy football” for the film business, as it assigns “quantitative scores to individuals, according to factors such as recent or past box-office performance or social media profile” (Rose). Scriptbooks approach is slightly different as its system uses script analysis as the basis for its prediction of a film's success, claiming to use “natural language processing” to determine the “commercial viability” of a scene and whether or not it should be promoted to certain audiences (Rose). Regardless of the nuanced differences between these programs, it is evident that they ultimately have a similar goal: to use AI’s ability to analyze past data — from box office failures, successes, and everything in between — to inform what kind of films are made in the future. 

The emergence of these softwares, and the film industry's interest in them, is no surprise. In the mid 2010’s, Ryan Kavanaugh sweet-talked his way into millions of dollars and the trust of major studios and entertainment companies due to one big promise: a computer program that could forecast a film’s success more accurately than any studio executive could (Wallace). Despite Kavanaugh’s concerning spending habits, questionable behavior, and lack of evidence, Hollywood seemed to have come rushing to his doorstep. In 2016, after various box office busts and accounts of money laundering and fraud, Kavanaugh’s empire came crumbling down along with the too good to be true premise it was built upon (Wallace). Kavanaugh's story and quick rise to success is demonstrative of a larger desire in Hollywood; to know something. While Kavanaugh’s software did not utilize AI technology and his erratic and fraudulent behavior was responsible for his demise, there are similarities between the promises of Kavanaugh’s software and the promises of AI companies like Cinelytic and Scriptbook. Cautionary tale or not, if there is one thing true, it's that the future is inherently unpredictable. 

Given the fact Cinelytic and Scriptbook’s systems are centered around a database of films and data from Hollywood’s past and are focused on predicting box office success, it is likely that star-studded blockbusters and big budget franchise films will be favored over riskier, more unique narratives. The issue of prioritizing profit over originality has already been an on-going issue in Hollywood. Box Office Mojo’s top lifetime grosses list is almost entirely franchise films, aside from stand out hits such as Titanic (1997), E.T. (1982), and Bohemian Rhapsody (2018). Since these softwares also analyze and suggest actors and directors based on their associated projected success levels, well-established creatives are more likely to be recommended by this AI. Furthermore, the issue of the historic biases of Hollywood and their films also comes to mind when considering how these systems utilize past films as data; it does not require an AI software to determine that white men have historically occupied positions of power both on and off the screen in Hollywood. During Ava Duvernay’s guest lecture for former Paramount CEO Jim Gianopulos’s course at USC this past fall, the filmmaker drew attention to the fact that minority filmmakers have to work twice as hard as their white counterparts as they “cannot afford to lose,” while directors such as Wes Anderson and David O. Russel are given a second chance even after their films flop (Duvernay). While there is no way to tell how the algorithms will process the inherent biases present in the industry and in their films, it is worth drawing attention to this potential issue given that Hollywood is still predominantly white and patriarchal.

A potential counter argument to my claim is that AI technologies could actually help diversify Hollywood and its content. By analyzing a multitude of relevant data and making this accessible to studios, AI softwares could potentially favor content and creatives that stand out from the masses as a means for greater success. Also, in making this data more easily accessible and referenceable to studios, AI could be used as more of a tool for comparison and research rather than a decision maker. 

Just as the future is unpredictable, it is impossible to predict how AI technology will impact the films of the studios that utilize it. However, in this case, I hope to be proven wrong in my argument that it will contribute to the production of unoriginal films. Regardless of the outcome, however, one thing is certain: these technologies are indicative of the formulaic future of Hollywood. Cineyltic and Scriptbook promote an algorithmic computer based approach to filmmaking, effectively reducing the humans involved on both ends of a film — the people who write, create, and act and the audiences that watch — to a data point. Netflix’s successful, albeit purposely low-key, use of AI and new found identity as a tech company has cemented the power of algorithms and data within Hollywood, as AI technology is the driving force behind their lucrative recommendation engine; an algorithm that the company assigned a value of $1 billion too (Ross). While technological advancements have been the driving force behind many major changes in the history of cinema — from the first ever motion picture to the technical evolutions that led to sound and color in film — the distinction between technology companies and film companies has been well defined throughout the decades. In today’s media landscape, however, this line has been blurred. And with studios like Warner Brothers partnering with A.I companies, it seems the line may get even blurrier. 

Works Cited

Fritz, Ben. The Big Picture, The Fight for the Future of Movies. Eamon Dolan Book, Houghton Mifflin Harcourt, 2018.

Bell, Breanna. Martin Scorsese Compares Marvel Movies to Theme Parks: ‘That’s Not Cinema.’ Variety. October 4, 2019. 

Debruge, Peter. With One Line, William Goldman Taught Hollywood Everything It Needed to Know. Variety. November 16th, 2018. 

Rose, Steve. ‘It’s a war between technology and a donkey’- how AI is shaking up Hollywood. The Guardian. January 16th, 2020. 

Wallace, Benjamin. The Epic Fail of Hollywood’s Hottest Algorithm. New York Magazine, Vulture. January 24, 2016. 

Dietrich, Conrad. Why Streamers Need Movie Theaters to Win. Medium. July 16th, 2021. Raevskiy, Mikhail. How AI Is Changing The Film Industry. Medium. October 7th, 2020. 

Duvernay, Ava. Diversity and Exclusion in Creative Content. CTCS 482: The Future of Creative Content. Guest Lecture at University of Southern California, October 10th, 2022.