LARGO AI REVEALS HOW AI CAN HELP LEVERAGE CASTING SUCCESS AT AFM 2021
MAN VS THE MACHINE
LARGO AI REVEALS HOW AI CAN HELP LEVERAGE CASTING SUCCESS AT AFM 2021
- Gender diversity in lead role representation has improved by 6% in the last three years
- Women rank within two of the top three matches to play next James Bond according to AI
- Idris Elba sweeps AI victory as audience’ favourite to play Bond vs AI’s favourite, Karl Urban
- Majority audience (53%) believe AI should be ethnicity blind
- August Wittgenstein revealed as best budget optimising match for new Rudolf Nureyev biopic, “Corsair”
Nov 4 Film and technology industry professionals took to the virtual stage at the American Film Market yesterday for a live and interactive webinar exploring ‘How to Cast the Right Actors for Your Film by Using AI’ revealing insights into
Hollywood’s most hotly debated topics, from ethnic and gender diversity – to the most fiercely speculated question; who will be the next 007?
Hosted at the American Film Market by CNN Anchor, Ana Maria Montero, alongside panelists; Producer and Talent Manager, Daryn Simons from Cohesive Entertainment Group, multi-award-winning Canadian film Director, Sean Cisterna, together with Largo.ai CEO, Sami Arpa, the interactive session demonstrated how AI can yield success in casting and ultimately Box Office revenue. The session was attended by film makers who were polled on some of the industry’s most contentious issues.
Gender diversity in lead role representation has improved by 6% in the last three years
AI provides the entertainment industry with a useful tool to challenge diversity problems and reduce gender and ethnicity bias. Overall, the data shows progress in both gender and ethnic diversity in movies over the past two decades.
Largo.ai aggregated the data from 22,000 movies over the last 20 years revealing that whilst there has been an improvement in diversity for gender within lead role representation – there is still a way to go. The data reveals that women being cast in lead roles has increased from an average of 32% in the last 20 years to 38% in the last three years. However, when considering second and third supporting roles, there is no change – revealing that the dominance of male casting still prevails. Digging deeper into genre specific insights, Action movies demonstrate the most significant progress within the last two decades, with lead roles given to women rising from 16% to 25%. However, 25% still represents poor gender diversity overall and that much work is yet to be done.
Man Vs The Machine: Should AI be ethnicity blind?
A game-changing factor in the battle to reduce society’s unconscious ethnicity bias in casting is that Largo AI’s software has been developed without providing the ethnicity of actors, making it blind to ethnicity. This enables casting directors to make informed decisions based on the best suitability and match for the role, regardless of ethnicity.
Because there is, arguably, an ongoing need to cast characters from a specific ethnicity in line with the story, Largo posed this question to its interactive audience and asked: Should AI be ethnicity blind? The results show an almost equal divide in opinion with a 53% majority vote in favor of omitting ethnicity as a consideration. This show that at this moment in time, there is no clear answer and the debate continues.
Women rank within two of the top three matches to play next James Bond
Taking film and literature’s most ubiquitous and iconic role of all time, Largo.ai then looked at the most hotly speculated question – who should step into Daniel Craig’s shoes as the next Bond? As an AI assisted casting process, the data reveals which actors are predicted to generate the highest grossing Box Office revenue.
Following Largo’s 2020 predictions, reported on by media worldwide, the AI provider this time broadened the search parameters to include international and female actors. Largo’s results revealed a 96.7% match for New Zealand actor Karl Urban with TV star Katee Sackhoff ranking as second favourtie with 94.4% – pipping fellow female contender, Paula Patton (92.4%) to the double-0 post. This is a significant result for the gender diversity debate. Henry Cavill ranking as 4th best matched at 92.3% still saw him rate higher than the top original Bond Roger Moore (91%).
Largo’s AI scoring system works by creating a DNA footprint of each actor which comprises over 1000 on-screen character attributes, such as character typology, story elements, and cinematographic tones. The system then analyses each of these attributes – weighing up audience reaction by cross-comparing the data against historic films. The systems also looks at the actual film character (James Bond rather than Sean Connery, the actor) and creates another DNA footprint for the character. It then cross compares the actor’s DNA footprint against the characters’. In this instance, the software looks for correlation between the two. This is crucial given research and previous projects have found a very high correlation between the success of a movie and the actor/character match rate – such correlation is more important than the popularity of the actor.
Taking a look back retrospectively at the debate getting most people’s tuxedos in a twist, Largo.ai also ran a search to create a definitive ranking of the best ever James Bond actors to date. Largo.ai’s deep dive showed Roger Moore taking the definitive 007 crown with 91.8%, followed by Pierce Brosnan with 90.8%, Sean Connery with 90.3% and Timothy Dalton trailing behind with 88.5%.
In contrast, the audience voted Idris Elba as the favourite casting choice for Bond, championing ethnic diversity. This was followed closely by Henry Cavill and Paula Patton who matches AI’s ranking within third place.
AI Assisted Casting To Maximise Budgets
Largo.ai then ran autobiographical film, Corsair, under the AI microscope to cross-compare casting considerations such as popularity, face similarity and AI Box Office predictions. Currently in pre-production, Corsair, is written and produced by Radik Kudoyarov and based on unknown facts of the life of Rudolf Nureyev – the legendary ballet dancer defected from the Soviet Union to the West in 1961, died at age 55 from AIDS. He went on to dance with The Royal Ballet in London and served as director of the Paris Opera Ballet.
Currently the most hotly contested actors in line for the lead role include Douglas Booth, Sam Heughan, Rupert Friend, Bill Skarsgard, and August Wittgenstein. Largo presented its AFM audience with best AI matches (best matched and highest grossing Box Office prediction) as well as popularity or ‘STARmetre’ according to IMDB.
According to Largo AI, the best financial yields were revealed as Rupert Friend with a 93% AI match, $152M Box Office prediction and STARmetre score of 1,627. August Wittgenstein is also a 93% AI match with $150M predicted Box Office revenue and a STARmeter score of 14,214. In this case, AI would suggest casting August Wittgenstein who has an equal AI match, similar Box Office yield but – crucially – his STARmetre IMDB rating (14,214) indicates he would be more affordable, commanding less of a fee, thereby maximising budget for this movie.
When the audience was asked (prior to AI data being revealed) they considered Rupert Friend as best matched for the role narrowly above August Wittgenstein – showing human instinct is very clearly aligned. Yet AI helped to distinguish how casting August Wittgenstein would ensure the same return but for less investment.
Given this is a biopic, a crucial factor to consider is how important facial similarity is when casting. According to Largo AI, Cillian Murphy, Tom Weilling, Harry Connick and Julian Sands are the most similar matches, representing the different ages within Rudolf’s lifetime.
Sami Arpa, Co-founder and CEO, Largo Films, said: “It is encouraging that the audience supports our idea of keeping AI ethnicity blind with 53% in favour. It’s also interesting to see that those human instincts, our audience, is very closely aligned with our James Bond and Corsair casting. AI has evolved rapidly in recent years and is now an integral part of successful filmmaking – casting is just one small element AI can support. It’s important to understand that the role of AI in filmmaking is best kept as an assistance tool – helping to quickly provide insights that underpin human instincts rather than replacing them. We’re proud to join the debate and open up the power and possibilities of data-driven storytelling.”
Largo is a Swiss company providing Largo.ai as a data-driven intelligence platform to the audio-visual industry. Largo.AI is able to analyse a film from as early as the script phase and provide insights also at post-production and distribution stages by analysing first video versions of the movie. The company was awarded the best start-up prize at the San Sebastian Film Festival 2019 and was selected as one of Horizon 2020 Top Start-ups in Berlinale. The company was selected by Slush as one of 15 finalist start-ups among 1500 start-ups all around the world. Largo recently raised CHF 750M from private investors and private programs, which brings a total financing of CHF 1.8M since its launch. The company has gained worldwide press coverage in publications such as Variety, Screen International, Indiewire, and 24 Heures.
Source: Rave PR
- VFX Supervisor Kyle Odermatt Talks the Making of ‘Raya and the Last Dragon’
- ftrack Studio has a ton of new features
- 13th Annual Kidscreen Awards Announces Winners for 2022
- Celebrating a decade of ftrack
- Unreal inSiggraph2022
- Smartphones top the rankings of the most widely used gaming platforms in Germany
- TOMY International Announces Leadership Changes
- Launch of Kloa: Child of the Forest on Kickstarter