AI-driven dialogue analysis confirms popular movies have grown more violent over the past 70 years

Movies often reflect the predominant societal and cultural values at the time they were shot. These values can be expressed in various elements of a film, including the interactions between characters, their communication styles and their characterizing traits.

Over the past few decades, some parents and scholars have expressed their concerns about the recent evolution of Hollywood Oscar-nominated and blockbuster movies, suggesting that they contain significantly more abusive and violent content than movies released during earlier historical periods. Yet, these debates are often grounded on a general perception of films as opposed to detailed analyses of films.

Two researchers at the University of New South Wales in Australia recently carried out a study aimed at better understanding the differences between the content and dialogues of movies released over the past 70 years, using a class of well-known computational models known as large language models (LLMs). Their findings, on the arXiv preprint server, confirm the hypothesis that movies have become more violent over time while also highlighting movie genres that appear to feature the most abusive and violent content.

“This study uses LLMs to explore the longitudinal abuse and sentiment analysis of Hollywood Oscar and blockbuster movie dialogues from 1950 to 2024,” wrote Rohitash Chandra and Guoxiang Ren in their paper. “By employing fine-tuned LLMs, we analyze subtitles for over a thousand movies categorized into four genres to examine the trends and shifts in emotional and abusive content over the past seven decades.”

The researchers used their LLMs, which were previously trained on datasets containing emotional and abusive texts, to analyze a total of 1026 movies released between 1950 and 2024. These included the top 10 blockbuster films for each year, as well as movies that were nominated for the Oscars.

They divided these movies into four broad genres, namely action (478), comedy (222), drama (285) and thriller (41), using data collected from the International Movie Database (IMBD) website. The LLMs they developed analyzed the dialogues of each movie, classifying the emotions conveyed as positive or negative while detecting how often different words were repeated.

The models were also trained to pick up general sentiments expressed in movies, including emotional features such as humor, sadness, denial, anger, anxiety, pessimism, optimism, and gratitude. The models’ analyses revealed that the dialogues of movies and their underlying sentiments have changed significantly over the past decades.

“Our findings reveal significant temporal changes in movie dialogues, which reflect broader social and cultural influences,” wrote Chandra and Ren. “Overall, the emotional tendencies in the films are diverse, and the detection of abusive content also exhibits significant fluctuations. The results show a gradual rise in abusive content in recent decades, reflecting social norms and regulatory policy changes.”

Overall, the results of this recent study appear to confirm the idea that popular movies have become increasingly violent and abusive over time, with some genres displaying the highest amount of violent content. Interestingly, however, the researchers found that many positive emotions are consistently expressed in movies of different ages, including humor and optimism.

“Genres such as thrillers still present a higher frequency of abusive content that emphasizes the ongoing narrative role of violence and conflict,” wrote the researchers.

“At the same time, underlying positive emotions such as humor and optimism remain prevalent in most of the movies. Furthermore, the gradual increase of abusive content in movie dialogues has been significant over the last two decades, where Oscar-nominated movies overtook the top ten blockbusters.”

The new insight gathered by Chandra and Ren could inspire further research into the social factors influencing the observed changes in movie dialogues, as well as the implications of these changes. Moreover, their approach to sentiment analysis could potentially be used to study the emotional qualities of a broader range of movies, as well as TV series, books, or even artworks.

© 2025 Science X Network