Can Artificial Intelligence Help With Movie Casting Decisions?

Hollywood loves remakes.

Thanks to a business model that thirsts for ever increasing revenue streams, the U.S movie industry is increasingly hedging its bets on remaking films that have already established some form of brand recognition with movie audiences.

While movie fans are split on the idea of seeing their favorite movies remade over and over again, the simple fact of the matter is that remakes make money.

How much money exactly?

This article from CNBC states that, “Disney’s remakes have made more than $7 billion globally since 2010.” The total figure for the entire U.S film industry is many times more.

Even more lucrative than the remake is the movie franchise, a series of movies featuring the same characters (and often very similar narratives), which lure audiences back time and time again to see their beloved heroes win the day.

The top 10 movie franchises have earned in excess of $50 billion USD, a figure that shows just how powerful their draw is to international audiences.

However, there is a problem that both the remake and the franchise must overcome if they are to be successful, namely finding the right actors to replace the ones that starred in the original, who might now be too old or have decided not to return for some other reason.

A good example of this challenge can be seen with the James Bond franchise, now in its 6th decade, which has seen a total of 7 actors playing the British superhero spy. Fans of the series are split about who is the best bond and who is the worst, a debate which highlights the enormous challenge that lays behind casting actors to replace ones who have left.

Casting failures can and have doomed movies, which is why casting directors need all the help they can get.

This is where artificial intelligence comes in.

To demonstrate the power of AI-assisted moviemaking to empower casting directors with the information they need to get decisions right, LargoAI has undertaken an analysis of 3 movies to find the most suitable actors for the lead roles.

The analysis includes actors who were previously considered for the role but were not chosen.

Let’s take a look at what it found.

Case Study 1: Batman

Over the last 25 years, no less than 5 different actors have played the character of Bruce Wayne, aka. Batman. While actors such as Michael Keaton and Christian Bale have made the character their own, others such as George Clooney have been dismal failures.

LargoAI was first asked to compare the suitability of the role for Christian Bale and the upcoming Batman, actor Robert Pattinson.

LargoAI was then asked to make alternate suggestions for the role including comparing actors who were previously considered for the role.

1990s James Bond, actor Pierce Brosnan, who was a commercially successful Bond, was also added to give some perspective.

Here are LargoAI’s results for Bruce Wayne:

1. Bruce Wayne/Batman:
Selected actors:
Robert Pattinson 72.5%
Christian Bale 83.2%
AI proposed actors:
Charlie Hunnam 87.2%
Sam Heughan 85.6%
Lucas Black 91.9%
Actors considered for the role previously:
Jake Gyllenhaal 77.6%
Keanu Reeves 81.2%
Ethan Hawke 73.8%
Johnny Depp 80.6%
Josh Brolin 80.7%
Pierce Brosnan 70.5%

Batman Results Analysis

The results for Christian Bale, who is generally regarded as one of the best Batmans, can be used as a marker with which to compare the results, as can those of Pierce Brosnan, who despite not ever playing Bruce Wayne, did play a similar character in an overlapping genre.

Most interestingly, LargoAI determined that upcoming Batman Robert Pattinson is one of the least suitable actors for the role, awarding him the second lowest score of 72.5%.

This is a very important prediction that can be used to validate the power and accuracy of AI moviemaking. The latest Batman is not due for release until October 1, 2021, which will be a full year after this article was published.

While the movie will make a profit, the success or failure of Robert Pattinson will be measured by whether or not he is lauded by fans and ultimately, whether he returns as Bruce Wayne a second time.

LargoAIs results suggest that he is less likely to return than most of the other actors analyzed.

 AI Proposed Actors:

LargoAI’s results for alternate actors certainly raised a few eyebrows.

Largo AI
largo AI

Charlie Hunnam, Sam Heughan, and Lucas Black where suggested by the system. Not only that, their suitability scores far exceeded those of Robert Pattinson.

Lucas Black, was a surprise suggestion for the role of Bruce Wayne. However, he is certainly someone who could carry off the role because of his proven ability in action movies as well as his on-screen charisma. Examining his past roles, he would almost certainly be able to give the character more panache and dark elements than Robert Pattinson.

Actors Previously Considered for the Role:

With the exception of Pierce Brosnan, the results are very close when it comes to actors who were rumored as being considered for the role of Bruce Wayne in the past.

While it is important to note that LargoAI is configured to establish patterns relating to contemporary audiences and not those of the past, the results are still interesting.

Highest scorers Keanu Reeves and Johnny Depp have both scored big in other action franchises. The exception, Josh Brolin, however, has not.

So why would LargoAI suggest an unproven franchise actor for this lead role?

The reason is that LargoAI doesn’t simply analyze similar criteria to establish its insights and suggestions. Rather, the system uses a large diversity of data sources, including its own past results, to calculate the most accurate results.

While Josh Brolin has not proven his ability to carry franchise, a number of factors suggest that he would make a good Batman.

Firstly, he has proven that he is able to bring a real depth to the roles that he has played, even when playing darker characters such as Thanos in the Avenger movies. He has also played a number of hero roles, some of which were complex brooding characters, similar to Bruce Wayne.

Brolin, like Christian Bale, is much loved by the critics as well as most film buffs, something which would certainly draw a wider cross section of people into seeing himas Batman. His acting talent is highly regarded, roles such as George W. Bush and Dan White in Gus Van Sant’s Milk (2008) have shown that he is able to draw the audience’s gaze, even when not in the lead role.

Josh Brolin would be an interesting Batman, and while maybe not the best, would certainly be more suitable in the role to the likes of Ben Affleck and Val Kilmer.

Let’s take a look at Batman’s supporting actors:

Selected actors:
Katie Holmes 67.1%
AI proposed actors:
Emma Watson 81.7%
Jennifer Lawrence 88.2%
Michelle Dockery 88.2%

LargoAI confirmed many critics’ unimpressed viewpoint in regards to the casting of Katie Holmes as Rachel in Christopher Nolan’s Batman Begins (2005). It found her to have a suitability score of 67.1%, and instead suggested Jennifer Lawrence, Michelle Dockery, and Emma Watson.

All 3 of these actresses could carry the role. However, LargoAI’s suggestion of Jennifer Lawrence seems like the best match. Her performance in The Hunger Games (2012) showed that she has the star power, acting talent, and onscreen charisma to make the role her own.

Selected actors:
Michael Caine 81.3%
AI proposed actors:
John Noble 90.2%
John Wesley Shipp 90.8%
Liam Neeson 87.3%

LargoAIs analysis of the most suitable actor for Alfred is quite surprising. Ever since Tim Burton’s Batman (1989), Alfred has been a British character, something which has lent an air of sophistication to the role.

The system awarded popular Alfred, Michael Caine, a score of 81.3%. Meanwhile, it suggested American actor John Wesley Shipp, giving him a higher suitability score of 90.8%.

Could an American Alfred succeed?

It is certain that John Wesley Shipp could play the role, whether Batman fans would accept the swap over is not. This would be up to the casting team to decide. He is certainly an interesting selection however.

Liam Neeson is an actor who would most certainly make a good Alfred. He is a powerful onscreen character who could easily provide the stern support and guidance that Alfred has always lent to the youthful Bruce Wayne.

Case Study 2: Mary Poppins Returns

When it was announced that there was to be a follow up Mary Poppins movie, many critics’ heads sank into their hands.

The film received a budget of $130,000,000 and Emily Blunt was cast as the lead. Despite receiving mixed reviews from the critics, Mary Poppins Returns was a commercial hit, earning $349,545,966 worldwide.

Let’s see what Largo AI thought of Emily Blunt and what other actors it found suitable for the role:

Mary Poppins
Selected actors:
Emily Blunt 90.7%
AI proposed actors:
Lily James 93.1%
Mandy Moore 87.8%

Mary Poppins Returns Results Analysis

While critics reported being underwhelmed by the songs in Mary Poppins Returns, most were impressed by Emily Blunt’s performance as Mary Poppins. Vanity Fair stated that “Blunt was precisely the right person to take Julie Andrews’s umbrella.”

LargoAI was in agreement with this opinion. Its suggestion of Lily James and Mandy Moore as possible candidates would be worth consideration too.

Selected actors:
Lin-Manuel Miranda 95.3%
AI proposed actors:
Zachary Levi 99.8%
Richard Madden 92.60%

Lin-Manuel Miranda’s performance as Jack also scored well with critics. LargoAI also scored him as highly suitable.

However, Zachary Levi’s suitability score of 99.8% is almost unprecedented considering that LargoAI cross references so many different data sets. Were the movie being cast today, this would certainly cause AI-assisted casting directors to seriously consider him.

Case Study 3: To Kill a Mockingbird

Based on Harper Lee’s Pulitzer Prize-winning novel of the same name, To Kill a Mockingbird (1962) is a movie classic. There have been rumors of a remake for a decades, however, no studio has yet taken this enormous challenge on.

While this movie classic might be better left untouched. Here are the results for LargoAI’s actor analysis:

1. Atticus
AI proposed actors:
Rupert Friend 93.3%
James Norton 93.7%
Ryan Gosling 91.0%

Largo’s suggestions for Atticus are interesting.

In To Kill a Mockingbird, Atticus Finch is almost 50 when he is asked to defend Tom Finch. Gregory Peck was 46 when he played the role, so it is not too much of a pinch that these 3 actors could play the part. However, two are British, while Ryan Gosling is Canadian, so all would have to master a difficult southern U.S accent in order to play the part.

LargoAI’s results can be seen to reflex the increasingly limited diversity of the type and look of actors playing lead roles by contemporary audiences. Both male and female leads need to conform to a narrow paradigm of beauty, what could be referred to as the ‘generic good looking male or female’, if audiences are going to accept them.

Sadly, the results seem to highlight that it is unlikely that an actor with Gregory Peck’s looks would ever be considered for the role of Atticus were he to audition today.

2. Scout
AI proposed actors:
Brooklynn Prince 100.0%
Jaynee-Lynne Kinchen 89.2%

Young Mary Badham gave an impeccable performance as Scout in To Kill a Mockingbird, for which she was nominated for an Oscar. Filling her shoes would be no easy task.

LargoAI suggested both Brooklynn Prince and Jaynee-Lynne Kinchen.

Brooklynn Prince scored a 100% suitability, which would almost certainly see her cast as Scout were the movie to be remade while she is still in the right age group.

Why Data-assisted Moviemaking Matters?

The movie industry needs all the help it can get to survive.

With global behavior patterns shifting in favor of interactive entertainment that is instantly available from smart phones and tablets, young people, in particular, are worryingly turning their backs on movie theatres as a way to watch movies. Many are even giving up on movies altogether.

The relatively steady international movie theatre audience numbers hide the truth about what is happening to national movie audiences. Increasingly, numbers are in decline as audiences lose interest in paying high prices to see the same old stories. This phenomenon has been exasperated by the Covid 19 pandemic.

It is only because movie audience numbers are growing internationally, as developing countries take up the slack, that this decline is being hidden.

The industry needs help to win audiences back. In order to do this, it needs to understand audiences better. This is where AI-assisted moviemaking comes in.

AI assisted moviemaking allows filmmakers to gain insights that help them to understand their audiences more. With such insights, they will have the power to improve their movies to ensure their success.

Understanding what audiences like and dislike is the key to future success. AI-assisted moviemaking is the key to this understanding.