AI to Foresee Blockbusters

20Th Century Fox 20Th Century Fox
Via 20th Century Fox

20th Century Fox is one of the most iconic filmmaking studios of our time but in its entirety, a business endeavor. This business endeavor has allegedly been looking for ways to predict the future of its films hence paving out ways for it to earn more. Artificial Intelligence is the epitome of technological advancements by it being capable of simulating real-life patterns in giving a more precise look into future probabilities.

It has been reported by a paper published by researchers in the company that they are looking into using machine learning on the contents of movie trailers. This practice is supposed to predict whether the general public and the audience will like the movie. The AI software analyzes the trailer footage frame to frame.

As the first step, objects and events are labeled in each of the viewed frames generating a frequency-oriented data. The final data analysis circulates the idea that a similar group of people will be attracted to watch a film with a similar type of objects and events in its trailer.

Filmmaking studios have already been after data of a similar type but with the use of traditional methods, including interviewing consumers or having questionnaires filled. However, dividing these audience groups further into micro-segments that inevitably comes with the use of AI would make the concept even more practical. The studios, therefore, need to know what people want to see, so they know what movies to make.

The paper says:

Understanding detailed audience compositi+on is important for movie studios that invest in stories of the uncertain commercial.

The most recent and exciting example presents itself in the form of using AI to stir up predictions concerning the 20th Century Fox superhero film, Logan. This film was released in 2017, and the test using AI was performed to check what other films would Logan fans want to watch. The ‘experimental movie attendance prediction and recommendation system’ was named Merlin.

Via 20th Century Fox

Also read: Predicting religious violence; the future of AI

Using Logan was tricky and smart on the part of 20th Century Fox since Logan is not only a popular superhero action film, but it is one with a dark and profound plot. The treatment of its main character ‘Wolverine’ by the film creators gives him depth and layers. The test for Merlin was to detect that successfully.

In its first step of smart deduction, the machine analyzed the trailer and kept count of objects and events as shown in the trailer. Unsurprisingly, Merlin also kept up with the fact that more slow-motion shots of the protagonist in the trailer hint to it being a drama.


Moreover, the use of fast-moving sequences in the trailer suggests it is an action thriller. The finalized data was a complex mix of acknowledgments of several shots of a brooding and bloodied protagonist as well as of this film having superhero action sequences. Merlin was programmed to select top twenty movies that Logan fans would like to watch as well.

You may also like: AI to oversee Google’s undersea whale analysis

The next step was comparing data collected from conventional methods like interviews with data that Merlin presented. Out of the top twenty, Merlin got eleven right, including all the top 5. It is interesting how it made correct guesses; the likes of John Wick Chapter 2 when it was not a superhero movie at all but had a similar brooding and bloodied protagonist. It is also equally impressive how Merlin could point to Tarzan and The Revenant as possible answers simply because all these trailers showed plenty of trees.

This story proves why a studio giant like 20th Century Fox would think of employing AI but also tells why computers are not reliably accurate yet.

Add a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Keep Up to Date with the Most Important News

By pressing the Subscribe button, you confirm that you have read and are agreeing to our Privacy Policy and Terms of Use