AI for the Film Industry: An Experimental Analysis of Its Possibilities

PillarStatus
Research Incomplete

 

 

Description

Is it possible to build a film industry based on artificial intelligence in Latin America? In Japan, it has been reported that many illustrators work under deplorable labor conditions. Could AI increase productivity in the creation of cinematic content and, therefore, improve working conditions? These were some of the questions we asked ourselves and that motivated us to undertake this research.

In this project, we experimented with videos generated by artificial intelligence to determine whether the current state of the technology already makes it possible to produce, at least, a short film. As part of this exploration, we tested a wide range of AI tools such as Midjourney, RunwayML, Krea, Kling, and MeshyAI, among others.

 

Incomplete Project

Carrying out social-impact projects is extremely challenging without a management team fully dedicated to their execution. Agenticus was founded with the goal of applying our knowledge to generate a positive impact on society. However, all of us involved in this initiative must also pursue other activities to sustain ourselves financially.

This limited availability of time prevented the completion of the research and, consequently, the publication of its findings. Even so, the process left valuable lessons about the challenges and opportunities involved in integrating artificial intelligence into film creation—particularly within the Latin American context.

 

 

Notes on What We Learned

1. Incorporating AI into Short Film Creation

For now, AI is a tool available to filmmakers. By itself, AI is not yet capable of creating a coherent or consistent short film — let alone a full-length movie — without human intervention. To illustrate the potential ways this technology can be used, we conducted a series of experiments as proof of concept. We focused our exploration on the creation of an animated short, since current AI generations more closely resemble animated films than those with human actors. Below are key stages in which AI can meaningfully participate:

  1. Keyframe Design: To animate a cut that usually lasts a few seconds, we propose using image-to-video models. The process begins with the creation of keyframes or reference frames. These can be generated with image-generation models; however, for now, the most efficient and precise method is still to design keyframes manually, either through illustration or photography.
  2. Interpolation with Image-to-Video Models: Using AI tools, the intermediate sequences between keyframes are generated. The technology is capable of interpreting the desired styles and transitions, creating animations or images that maintain the aesthetic coherence established by the director. This process accelerates production and enables the exploration of variations that might not have been imagined in a purely manual workflow.
  3. Voice Generation and Automatic Translation: AI-generated voices have advanced to the point that, in many cases, it is difficult to distinguish a synthetic voice from a human recording. This allows a single creator to bring multiple characters to life, each with distinct voices, languages, or accents—opening narrative possibilities once reserved for large productions. Moreover, fine adjustments in tone, rhythm, and emotion can be made without re-recording, offering an unprecedented level of control.

 

2. AItist: The New Kind of Artist Who Integrates AI into the Creative Process

Throughout the history of art, every technological innovation has initially been met with skepticism before being accepted as a legitimate form of expression. Photography, for example, was once considered a mechanical technique devoid of artistic merit—until its creative potential was recognized. Similarly, conceptual art, which emphasizes the idea over execution, was also debated before being fully acknowledged. Even video games, long viewed merely as entertainment, are now discussed as an art form because of their narrative and visual depth.

AI-assisted artistic creation now stands at that same turning point. For many, works generated with AI lack authenticity because they do not stem from human intention or emotion—a fair criticism, since the only human element in those cases may be the prompt itself. However, we believe that if there is a human in the loop, then a work, even if partially generated by AI, is an authentic piece of art created by an AItist (an artist who uses AI-generated content—AIGC—as part of their creative process).

AI does not replace creativity; it amplifies and transforms it. In this new paradigm, the question is not whether AI-generated content is art, but how we redefine art in light of these emerging tools. Just as the camera did not eliminate the painter, AI does not eliminate the filmmaker—it opens the door to a new generation of creators with expanded languages and possibilities.

 

3. The Future of Cinema and Streaming

The growing accessibility of AI tools is driving a profound transformation in the audiovisual industry. What once required complex technical teams, large budgets, and months of production can now be achieved by a single person with a computer and the right knowledge. This democratization of audiovisual production will inevitably lead to an explosion of content—short films, movies, series, and animations created by individuals or small teams across the world.

This unprecedented phenomenon in the history of cinema will directly affect how content is consumed. The overwhelming abundance of works will make the traditional monthly subscription model for closed libraries increasingly unsustainable. In this context, two possible scenarios for the future of streaming emerge:

First theory: Pay-per-view model
Instead of paying a monthly fee, users would pay only for the content they actually want to watch. This model would allow audiences to curate their own viewing experience and could favor niche or independent productions. Platforms such as Vimeo On Demand or certain digital premieres already use this approach, which could become widespread as AI-generated works multiply.

Second theory: A return to the advertising model
Another possible outcome is a model similar to traditional broadcast television or YouTube: users can access content freely but must watch ads in exchange. Those who wish to avoid advertising could pay for a premium subscription. This system, already common in services like Spotify and YouTube, could extend to films and series through new platforms specializing in AI-generated content.

The future of cinema will be increasingly personalized, fragmented, and dynamic. Audiences will no longer depend on major studios to access new stories. Instead, millions of creators, assisted by AI, will populate the audiovisual ecosystem with diverse proposals. The main challenge will lie in visibility, curation, and sustainability—where quality will have to stand out amid the sheer abundance of content.

 

4. Final Thoughts

  1. The emergence of artificial intelligence in audiovisual creation marks a turning point in the history of cinema. What was once reserved for studios with large budgets is now within reach of anyone with an idea, a computer, and the right tools.
  2. Creating a short film with AI does not mean eliminating the human element; on the contrary, it means assuming a new role as director, curator, and editor of automated processes. Human creativity remains the driving force behind every meaningful work—even when it relies on algorithms to materialize its vision.
  3. Cinema, like all art, is a way of interpreting and giving meaning to the world. In this new landscape, the creator evolves: they must not only learn to write scripts or operate cameras but also to dialogue with generative models, design effective prompts, and make aesthetic decisions. The AItist thus emerges as a new figure born from this technological and artistic revolution.
  4. Audiovisual distribution is also undergoing profound change. Just as AI is democratizing production, technology is opening paths toward more direct, decentralized, and personalized distribution models. Systems may arise where each audiovisual work is consumed on demand through micropayments, donations, or patronage. We may also see hybrid experiences that combine free, ad-supported content with membership models centered around filmmakers or collectives—just as in other creative disciplines.

 

Volunteers: