Connect with us


Should We Use Technology to Predict A Film’s Success?

Warner Brothers recently signed a deal with LA startup Cinelytic, which, according to its website, “supports studios and independent content companies to make faster and better informed greenlight, acquisition, and release decisions”. The studio’s primary goal is to improve “marketing and distribution” practices through the use of artificial intelligence. Cinelytic allows users to select popular actors and returns an estimate of the film’s performance at the box office. This way, Warner Brothers could create the “perfect” movie – or at least perfect by industry standards.

It’s no secret that the film industry tends to follow a certain mold when it comes to producing and consuming content. The lack of diversity in award nominations, writers’ rooms, and casting is generally excused by the assertion that only the best people were selected or that the decision was influenced by what they believe the public wants to see. Even people and organizations that aim to highlight those traditionally underrepresented in film show pitfalls in other areas.

Where bias poses an issue, it makes sense to bring in a neutral entity, and in today’s society, that neutral entity is often technology. However, as recent reports have shown, artificial intelligence is not neutral in the slightest. The viral project ImageNet Roulette, which returned captions based on images uploaded by users, showed that technology often reflected the negative and prejudiced views of its consumers. Facebook uses machine learning to target ads to its users based on what advertisers believe certain demographics should see or what they believe consumers will do. Technology is ultimately developed by people, so its “neutral” will reflect what its creators perceive to be neutral.

In the case of Cinelytic, film studios that use it will only continue to pursue the kinds of stories that they already believed to be successful without external consultation. Yet, the industry has repeatedly shown that they don’t actually know what audiences want – for instance, Bombshell (2019), the film about female reporters at Fox News confronting Roger Ailes’ harassment, was criticized for telling a story that didn’t appear to know its target audience. Aside from general public opinion about the subjects of the film, critics pointed out that there was nothing new about the story being told, and even an all-star cast was not enough to make it a hit.

Making different combinations of popular actors to predict a film’s success seems more like a fun game than an actual strategy. Given that users have already said the program didn’t do much to change their minds about the films being produced, it’s evident that the industry will continue pursuing what they already see as lucrative without Cinelytic’s help. It’s also concerning that film studios still emphasize commercial success above all else; while wanting to make a profit is justified, it seems that the industry is more interested in capitalizing on trends than actually understanding the public’s desires.

Likewise, commercial success is often held up as a reason to eschew fair representation or other important aspects of a story. Rumors about avoiding Chinese censorship have impacted the Star Wars fandom for months, although the most recent release showed that it was not an issue. There have been numerous stories about aspiring filmmakers being shot down because “movies about minorities don’t sell”. Beyond that, one would have to consider the metrics of who’s considered a more popular or worthy actor; what about those who are just starting out in the field? How are those who have been blacklisted impacted? There are far more questions than answers for an initiative that is meant to respond to preexisting questions.

What is most troubling about the use of AI in film is the assumption that deciding what audiences want is as easy as following a simple formula, when in reality, we’ve been asking for a departure from the formulaic for years now. The question is not whether using AI will work, but rather if we should use it at all.

Featured image via Deadline

Voted Thanks!
Nadia Bey
Written By

Nadia is a student journalist from North Carolina and the current Books Editor for Affinity. In addition to reading, she is interested in science, pop culture and policy.

Click to comment

Leave a Reply

Your email address will not be published.

Trending Posts


Copyright © 2018 Affinity Magazine