×
UC Berkeley study uses AI to confirm rising Hollywood diversity
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

AI-powered study confirms Hollywood’s increasing diversity: Researchers at UC Berkeley have used facial recognition technology to analyze on-screen representation in over 2,300 films, revealing a trend towards greater diversity in Hollywood since 2010.

  • The study, published in the Proceedings of the National Academy of Sciences, examined 4,412 hours of footage from both popular and prestige films released between 1980 and 2022.
  • Researchers found increased representation for women, Black, Hispanic/Latino, East Asian, and South Asian actors, particularly after 2010.
  • The diversity increase is not limited to a few films with all non-white casts but is evident across individual movies as well.

Technological and legal breakthroughs: The study was made possible by recent changes in copyright regulations and advancements in computer vision technology.

  • A new federal regulation eased “digital locks” on DVDs, allowing researchers to bypass copyright protections for scholarly purposes.
  • The UC Berkeley team had to purchase their own copies of all 2,307 films studied and use secure data handling measures to comply with strict usage rules.
  • The university’s Secure Research Data and Compute (SRDC) platform played a crucial role in enabling the research while maintaining data security.

Methodology and findings: The researchers used computer vision to track actors’ appearances on screen but relied on human perception for demographic categorization.

  • Actor demographics were determined using Wikidata for gender information and user surveys for perceived race/ethnicity.
  • The study found that women’s on-screen presence increased from 25% between 1980-2010 to around 40% in 2022.
  • All groups except white men remain underrepresented in leading roles compared to non-leading roles.
  • Black actors were found to be underrepresented in award-nominated films compared to popular films, primarily due to disparities from 1980 to 2010.

Implications for future research: The study’s methodology and findings open new avenues for large-scale analysis of representation in film.

  • The research team has released non-copyrighted elements of their data to support reproducibility and further studies.
  • Future research aims to explore more nuanced questions about representation, including how actors are depicted and potential stereotypes or biases.
  • The automated approach allows for analysis of a broader range of films and more detailed questions about representation than previously possible with manual methods.

Broader impact on film studies: This research demonstrates the potential of AI and computer vision in accelerating cultural analytics of film.

  • The study complements traditional human viewing methods by enabling large-scale, granular analysis of representation trends.
  • Researchers hope to collaborate with movie studios and film scholars to further explore and understand representation in cinema.
  • This approach could lead to more comprehensive and data-driven insights into the evolution of diversity in Hollywood and its cultural implications.
With the help of AI, UC Berkeley researchers confirm Hollywood is getting more diverse

Recent News

AI-powered computers are adding more time to workers’ tasks, but there’s a catch

Early AI PC adopters report spending more time on tasks than traditional computer users, signaling growing pains in the technology's implementation.

The global bootcamp that teaches intensive AI safety programming classes

Global bootcamp program trains next wave of AI safety professionals through intensive 10-day courses funded by Open Philanthropy.

‘Anti-scale’ and how to save journalism in an automated world

Struggling news organizations seek to balance AI adoption with growing public distrust, as the industry pivots toward community-focused journalism over content volume.