Member-supported news for Southern California
Play Live Radio
Next Up:
Available On Air Stations
Support for LAist comes from:

GD-IQ: The new software that breaks down representation in TV and films

Ways to Subscribe

The Geena Davis institute has partnered with USC and Google to create a software that could give movies and television a "diversity score."

We hear all the time how Hollywood gets things wrong when it comes to diversity. So, how do you get the entertainment industry to get it right? 

Score them. 

That's the idea behind new software being unveiled Thursday at the Global Symposium on Gender in Media.

GD IQ - short for Geena Davis Inclusion Quotient -  measures things like screen time and speaking lines for various characters.

Take Two's Alex Cohen recently interviewed the minds behind this project:

  • Geena Davis, Founder and Chair of the Geena Davis Institute on Gender in Media 
  • Shri Narayanan, of USC's Viterbi school of engineering and technical research lead
  • Julie Ann Crommett, Google's Entertainment industry educator in chief

Interview Highlights

How did this first come about?

Geena Davis:  "My very heightened interest with how women are portrayed on screen started when my daughter was a toddler and I immediately realized 'Wow, there's far more male characters than female characters in what we're showing kids. I had played some really great parts in my career, I was aware of what adults were seeing, but I had no clue that in the 21st century, children's media would be so profoundly skewed.

So, I mentioned it around town to a bunch of colleagues, producers and directors and studio executives and every single person said 'No, no, no. That's not a problem anymore. That's been fixed.' So it was then that I became a data head. I decided that I wanted the numbers and once I had that if I was proved to be right I could go back to everybody and share it with them and that's pretty much exactly what happened."

Take me through the process. What happens when you plug-in a movie or film through this software?

Shri Narayanan: "For example just take the visuals, so shot by shot this computer algorithm goes through it and figures out, where are the faces? Once that's done, it tracks them through the scenes and then automatically another algorithm figures out what is their gender and then you can run through the data in real time...and so by the end of that you have specific numbers and in a quick way have how much screen time a person has and how it's distributed across various genders. Likewise, you can process the audio which is like a separate stream of information of figuring out who's talking when and for how long. We can process the music that goes along with it and the words that they're speaking."

And in the end as I understand it there will be a score, so to speak?

Shri Narayanan: "You can look at the raw scores, if you want to know how much screen time males vs. females got, we got that number. How much speaking time and so on. But you also want to put them together with the more sophisticated statistical sort of model to come up with the inclusion quotient that we've been talking about. Sure, we can come up with a number that kind of can summarize what's happening on screen."

There's quantity and then there's also quality. How do you get at that part of it when it comes to how diversity is represented?

Shri: "We can look at some of the psychological meaning of the words that are being put in the mouth of the characters and then with all the data in front of us we can begin to see. For example, in that analysis we found the divide between female characters, the voice that they were given and the male characters. Pretty distinct actually. These kinds of things are possible and our team is thinking about these directions to see how we can start to understand and analyze this content but also how to use it as a tool for discovery."

Julie Ann: "And I think Alex, something to add to what Shri and Geena have said as well. I think this tool is a powerful one also to celebrate the people who are actually doing a great job. I think that needs to be said as well, not everybody is actually sucking at this. We need to honor those people. I think there are actually some shows and some pieces of content that are doing a great job and we'll be able to highlight them through this content and that has to be celebrated as well because they may not be getting their due in some ways."

To hear the full segment, click the blue play button above.

Answers have been edited for clarity.

For the full study, see below:

Stay Connected