Subscribe now to Folha and get unlimited access to real-time news, 150 columns and exclusive newsletters! Exclusive offer, don’t miss out!
SÃO PAULO, SP (FOLHAPRESS) – There was a time when MGM was essentially an entertainment company. Today, in the hands of Amazon, the studio is a good example of the visceral connection between Hollywood and Silicon Valley, illustrating the growing importance of new technologies in the life cycle of a film.
Not in vain, artificial intelligence has spread throughout the film industry. While many were concerned about its use in screenplays and replacing actors, two main reasons for the strikes that paralyzed Hollywood two years ago, major studios focused their efforts on post-production.
Unions from both groups managed to block studios from writing scripts with robots or cloning actors without their consent—although there is a loophole allowing productions to be populated by synthetic extras. Other categories, with less political power, however, were left uncovered.
Today, Hollywood’s elite have normalized the use of AI to adjust special effects. Its use has also been experimented with when drawing conceptual art for sets and costumes, sound mixing, and creating storyboards – a sequence of illustrations that help guide a story that will still be filmed.
There are even debates about using technology to adapt movies and series for a specific audience after they are completed, in just a few hours. Imagine a plot with a spicy scene that can be re-edited to obtain a lower age rating and be shown on free-to-air television. Or consider a sequence between gay characters, excluded from a version released in Middle Eastern countries without affecting the flow of the story.
“It’s a facilitator for censorship. AI won’t change the core of the story, but it will make it more superficial,” says Marco Aurélio Casson, a professor of animation and cinema at Faap, who is currently developing a game with the help of this technology.
He noticed that many characters belonging to minority groups in recent games are undergoing a process of sterilization and homogenization, in order to fill a certain type of quota, without necessarily giving depth to LGBTQIA+ issues, among others. “AI comes into play here, fulfilling the need for a quick, almost immediate result, based on a limited reference database.”
For Casson, the debate on AI is complex because, on one hand, it is undeniable that there is creative constraint and job precarization. On the other hand, it accelerates and automates processes, allowing efforts to focus on the less mechanical parts of film or game production, and democratizes access to tools for creating an artistic work.
“I don’t see AI in a very different way from other earthquakes of the past. When motion capture techniques [a process in which a character is animated from the capture of an actor’s movements] emerged, there was this fear, and I didn’t see any animator positions being lost because of it,” says the professor.
In recent months, James Cameron, director of “Avatar” and “Titanic,” and Darren Aronofsky, of “Black Swan,” have made significant steps toward using AI. The first joined the board of Stability AI, and the second will produce short films with DeepMind, from Google.
Throughout my career, I have sought out emerging technologies to test the limits of what is possible in storytelling,” Cameron said in a statement sent to the press. “The intersection between AI and CGI [computer-generated images] is the next wave, and it will allow artists to tell stories in ways never imagined.
Lionsgate, in turn, has been doing business with Runway, the company that had access to the producer’s archive, which includes franchises such as “The Hunger Games,” “Twilight,” and “John Wick,” to train its image generation model. Before that, it was sued by an artists’ collective for feeding its robot with images protected by copyright. The case is still ongoing in court.
One of the contracts between Lionsgate and Runway included the development of a movie trailer that was not yet in production, based on a script, so the idea could be presented at a film festival. Thus, the debate about the use of AI is not limited to time and budget cuts, but also includes processes and tasks that would not exist without the technology.
For now, there are no concrete indications of job cuts in film sets or editing islands due to this new reality. What frequently happens, as workers in the industry report on social media and interviews, is the use of AI as support for human work.
An illustrator in Hollywood, who has worked on franchises such as “The Matrix” and “Transformers,” Reid Southen has been causing a stir on X, the former Twitter, with posts on the subject. He conducted an informal survey on the platform with 800 professionals in the field. Nearly half responded that they had been instructed to retouch images created by AI or use them as a reference in their work.
In other publications, he compares movie frames created by professional artists and special effects experts, like himself, to images generated by intelligences such as Midjourney. The results reveal clear copyright violations, so much so that Disney and Universal have taken the platform to court, which they consider “an endless well of plagiarism,” highlighting the delicate line on which studios have been balancing.
In the last race for the Oscar for best film, “The Brutalist,” distributed by Universal, generated controversy for using AI to adjust the accents of its actors and generate conceptual images of brutalist buildings, which were later redrawn by humans.
“Emilia Pérez” also came under scrutiny for refining the dialogue of its two female leads—Mexican in the script, but Spanish and American in the cast. Meanwhile, “Dune: Part Two” automatically colored the eyes of some characters. In all these cases, the term “artificial intelligence” was avoided in interviews and promotional materials.
There are many situations [in the industry] where the term ‘machine learning’ or any other is used, and in fact, it’s all AI,” actress Natasha Lyonne told the American magazine Vulture, at the opening of her studio Asteria Film Co., which she founded with her boyfriend, businessman Bryn Mooser, and which she says is the first in Hollywood to use AI in an ethical way. “It’s better to get your hands dirty right away than to pretend nothing is happening.
A Asteria, say its founders, aims to serve as public relations for AI, help calm tensions, and have artists take control of the machines, not tech industry billionaires. One of its projects involved commissioning 60 illustrations from an illustrator. They were used to train software that, once capable of replicating his style, could generate hundreds of new images and variations for a music video.
In Brazil, despite the delay, the story is not very different. Last month, the country gained its first AI specialized post-production studio. In Rio de Janeiro, NKanda 360 aims to improve the finishing quality of films, series, and commercials produced here. Its first project, “Lovers,” is being co-produced by Colombia.
“The process should be supported by AI, but not replaced by it. It is a new form of expression that is here to expand, not to replace,” says Carien Bastos, who founded the company together with Fernanda Thurann, an actress and producer from Brisa Filmes.
For them, AI can help the Brazilian audiovisual industry develop in areas where it has historically lagged, such as sound mixing, and expand the scope of its production, embracing genres in which there is little tradition due to technical and budgetary limitations, such as science fiction and fantasy.
Imagine the cost of creating a fire with practical effects. There are a large number of people involved—firefighters, paramedics, protective equipment. Then there are the risks. For a scene like this, lasting one and a half minutes, we already spend US$ 40,000. With AI, this cost drops, maybe to around US$ 5,000,” says Thurann. “It’s a technology that allows creativity to fly higher, especially because in Brazil we work with tight budgets.
But beyond ethical, legal issues, and the fear of strikes, another obstacle to large-scale adoption, here and abroad, is technical quality. Current tools still do not deliver images with the same technical precision as human intelligence, although technology is advancing rapidly.
In the same article from the Vulture magazine already mentioned, a special effects expert who did not want to be identified compared the experience of watching an AI-generated scene to tasting wine — only a true movie connoisseur will notice the differences in quality.
About The Author
