From November 5 to 7, 2024, when the Federation of Latin American Audiovisual Authors' Societies (FESAAL) and the International Confederation of Audiovisual Authors (AVACI) hold their annual congresses in Mexico City, the use of Artificial Intelligence (AI) in the audiovisual industry will be a necessary topic of debate.
At this point, it is undeniable that AI is rapidly revolutionizing the audiovisual industry. Those leading this transition are not only directors and screenwriters but also developers who see technology as a tool for innovation.
Christoph Behl, film director and developer, is one of them. Behl has managed to merge traditional cinematic storytelling with the potential of AI-based tools, opening up a fascinating debate about authorship, rights, and the future of cinema.
In this regard, the director will share his insights during the AVACI Congress, where he will deliver a lecture on “The Impact of New Technologies—such as AI—on the Profession of Screenwriters and Directors.”
An Innovative Project: "Property"
One of the clearest examples of how AI can shape audiovisual creation is Behl's upcoming project, a film titled Property, which explores transhumanism through the story of a man who, after his death, transfers his consciousness to the digital world and sues his daughter for control of his companies.
According to Behl, “the movie will be made 70 to 80% with AI tools,” allowing him to not only develop a futuristic story but also transform traditional creation processes.
AI Writing: The End of Screenwriters?
The impact of AI on writing is one of the most debated aspects. Behl takes a technical and pragmatic approach to using large language models (LLMs) like Chat GPT and Claude in screenplay writing. The key, he says, is to work with systems that allow the script to be structured in fragments, defining specific elements such as characters, locations, and the emotional tone of the film.
“One prompt might be: 'Write the first act of the script considering the synopsis, the treatment, the characters, and the descriptions of locations and music,’” Behl explains. This fragmentation of the script allows writers to collaborate with AI and optimize storytelling in a way that would be difficult to achieve without these tools.
However, Behl warns that current models still have limitations, such as the size of context windows, which restrict the amount of information AI can process and remember from previous conversations. "Don't try to write entire scripts with this," he suggests. For now, the optimal use is to break the process into pieces and leverage AI to improve coherence across different parts of the script.
Image and Video: Creating Worlds with AI
The development of images and video through AI is perhaps one of the most impressive, though also the most challenging, fields. Behl highlights three main approaches to video generation: text-to-video, video-to-video, and image-to-video.
In his own work, Behl uses tools like Runway and Kling to generate images and videos, though he points out that it is still difficult to achieve consistency in all elements of a scene, from lighting to camera movement.
Still, with tools like Flux, which he calls “the best image model out there,” Behl has been able to generate detailed images that are then animated using AI.
A key aspect of this process is the use of systems like ComfyUI, which allow for precise control over visual creation. Behl describes how he uses “LORA, a fine-tuned model for a single character,” to create consistency between the generated images. The workflow is detailed and technical, but what stands out is AI’s ability to create images that were previously only possible with the meticulous work of specialized human teams.
"All the images in this gallery were generated with LETZ.AI - Flux Model."
Sound: Voices that Act
In the realm of sound, Behl highlights tools like Eleven Labs, which allows for the creation of ultra-realistic artificial voices. But the advancement that excites him the most is speech to speech technology, which not only generates voices but also gives them the ability to act. “I act: 'Pleeeease, ice creeeeam!' And we made the machine act,” Behl explains, demonstrating that, while technology doesn’t fully replace human actors, it is already capable of convincingly imitating emotions and performance.
The Battle for Author’s Rights
The use of AI in audiovisuals is not without controversy. Behl stresses the importance of addressing the issue of author’s rights seriously. Currently, many AI models have been trained on data sourced from the internet without verifying the origin of the materials, leading to issues regarding the misuse of works protected by copyright.
Behl proposes a reverse-engineering approach similar to what the New York Times is implementing, to detect possible infringements and ensure that author’s rights are respected.
"The discussion is nothing more than a struggle for rights; images, movies, video games will be made anyway, with or without authors getting paid for their authorship," Behl asserts. Without clear legislation protecting creators, big tech companies could monopolize the authorship of works created with AI, leaving artists out of the compensation process for their work.
A New Golden Age for Audiovisuals
Despite the challenges, Behl is optimistic about the future of audiovisual creation with AI. “We are facing a golden age of fiction,” he says, noting that new tools will democratize audiovisual content production. Small teams of one or two people will be able to create films and tell stories at a level of quality that was previously only within the reach of large production companies.
However, he also warns of the risk that these tools, if not properly regulated, could lead to an even greater concentration of power in a few hands.
“It’s essential that [author’s rights] remain ours, that we understand AI is a tool in service, and that we don’t fall into discussions of whether it’s AI or human,” Behl concludes, emphasizing the importance of creators maintaining control of their works in this new technological era.
Artificial intelligence offers immense possibilities for audiovisual creation, from screenwriting to image and sound generation. However, as Christoph Behl points out, it also raises important questions about authorship and creators' rights in a world where machines are increasingly involved in the creative process.
The future of AI in audiovisuals is full of possibilities, but it will depend on the industry and lawmakers to define whether that future will be a utopia of collaboration and creativity, or a dystopia of monopolies and loss of rights.
Comments