Artificially intelligent gender measure
confirms gender inequality in film

Gender disparities in the film industry are under heavy debate. Studies show that on average there are twice as many male as female characters in films — similar ratios appear for the gender of the main protagonist, speaking lines, and the gender of directors.

At ScriptBook, we developed an automated tool to measure gender bias using deep learning to analyse screenplays. We measure: the amount of dialogue and overall presence devoted to male or female characters; the amount of interactions between them; and whether characters conform to gender stereotypes based on their actions. The final metric is the Bechdel Test — a movie passes if it contains at least 2 named female characters, who have a conversation with each other, about something other than a man.

Bechdel test results

Applying ScriptBook's system to thousands of scripts, only 38.5% pass the Bechdel test. By genre, romance and horror score much better than adventure, crime, or animation. And tracking over time, around 1980 only 25% of movies passed. By 2017 this had risen to 43% — slow but real progress.

The Bechdel test and beyond

The Bechdel test is iconic but limited. ScriptBook developed additional metrics: inter- and intra-gender interaction counts; a "stereotypicality" measure for whether character actions conform to gender stereotypes; character counts; and spoken lines per gender.

Use case — Mudbound (2017)

Mudbound passes the Bechdel test — but only for 2 scenes. Our additional metrics tell a different story: almost 75% of speaking lines are from male characters, and male presence reaches 71.5%. The stereotypicality metric places Mudbound among the most stereotypical scripts. Passing the Bechdel test is not sufficient to conclude gender equality.

Use case — Still Alice (2014)

Still Alice has more female than male characters, a female presence of 74%, and female characters who drive the majority of interactions and dialogue. Its stereotypicality score is among the lowest — female characters display exactly the kind of control, intent, and decision-making usually reserved for male characters.

Awareness has impact

Artificial intelligence provides a fully objective view on subjects where we might have developed unconscious bias. By exposing ourselves to objective metrics, we become aware of gender bias and able to address it. Implementing AI to reduce unconscious bias will hopefully lead to a much fairer distribution of roles to women in filmed entertainment.