Researchers from the University of Granada and the University of Cadiz in Spain created the first computer system with artificial intelligence (AI) techniques that could help scriptwriters devise storylines that have a good chance at blockbuster success.

The system does not write scripts automatically, although the team believes that one day it may be able to. Rather, it is a tool to help film scriptwriters and directors during the creative process.

The new system analyzes tropes in existing films. Tropes are commonplace, predictable and necessary clichés that show up repeatedly in films which allow directors to convey scenarios that are easy for the users to recognize.

The team created a methodology that understands how tropes work. It analyzes how tropes are related to one another in different genres to find which combinations may be the most creatively successful. Pablo García-Sánchez, a researcher at the UGR's Department of Computer Architecture and Technology and lead author of this study. Source: University Of Granada

To create the algorithm, the team consulted an online database called TVTropes which includes more than 25,000 tropes associated with 10,766 films and is constantly updated by fans. They used a free software called TropeScraper. The algorithm scraped and extracted a list of tropes used in the films.

A mapping exercise was conducted based on user ratings and popularity for each film according to IMDB. Ratings given by users depend on a combination of actions based on tropes, the interrelation between the tropes, actions taken based on the tropes and which films the tropes appear in. This allowed the team to gather information on the level of interest given in a storyline.

Network analysis of tropes was conducted using programmed algorithms to reveal the relationship between films that share similar tropes. This builds a picture of existing communities of tropes and communities of films. With this method, the team could measure the popularity of tropes and determine if they were transversal across all films or highly specific, specialized. They also could determine if a given trope is on the rise or on the decline.

The team plans to analyze areas of the 'troposphere' that receive more attention, are currently enjoying success and areas that are causing a decline in interest. They hope to study the evolution of tropes in a given genre, country or decade to better understand audience consumption and the public’s interaction with a given film.

The study was published in PLOS ONE.