What if, when booking a movie ticket, you had to choose not just the language or format, but also between ‘Regular’ and ‘Organic’? The difference? One is made entirely using Artificial Intelligence (AI), while the latter is created by humans. This was the premise of a global social media content template titled ‘Watching a Movie in 2050’, which imagined a future where audiences might have to make precisely this choice. While AI remains a polarising topic in the arts, it is only fair to acknowledge that the technology has opened up an industry that was once a far-fetched dream for many filmmakers and creators.
For 61-year-old Shane Sadanandan from Kerala, AI offered a way to bring a filmmaking dream that had remained unfulfilled for two decades to life. “I always wanted to make films. In the past two decades, I wrote about 10 scripts and approached several producers and industry insiders. But the doors never opened for me. So when AI became available, I thought, why not?” Shane says. In April, he began experimenting with Google Gemini, right from the basics of prompt-writing. Six months later, his entirely AI-generated feature film, Googilil Illaatha Oridam, is now ready for theatrical release.
The cost of creation While studios across the world are experimenting with AI to create realistic films and series, Azeem Kattali, co-founder and director of Eunoians Studio, sees both possibilities and pitfalls in the technology. “As an artist, I am worried about how AI has been ignoring copyright laws and intellectual property rights to create images and videos,” he says. “The machine converts existing art to data for reassembling it for your use. And the original artists get no royalty or credit. In a way, their works are simply stolen.”
Muthu Karthikeyan, a Chennai-based engineer who works in Dubai, says, “AI gave me an opportunity to bring ambitious ideas to the screen with the resources available to me. LUNAVOM (his sci-fi film series) grew from that opportunity.”
A similar journey led AI filmmaker and creative director Vigneshkumar, from Theni district, to create films such as God Mother, Comfort Kills, and Arakkan. Having worked as a professional video editor, he was initially apprehensive about entering the film industry, given its “stress, financial uncertainties, long production processes, and politics.” Generative AI, which he first explored as an editing tool, gradually opened up a new avenue for visual storytelling. “I don’t see AI as a shortcut to becoming a filmmaker. I see it as a doorway that allowed someone like me, who had always dreamed about making films but was afraid to pursue them, to finally tell the stories he wanted to. It gave me the courage and the means to pursue a dream I had kept aside for years,” he says.
Where’s the art?
When a machine generates the visuals, where does the filmmaker’s artistry lie? “I believe the art lies in the filmmaker’s intention and the decisions that carry it through to the screen. Why this character? Why this moment? What should the audience feel?” Muthu says, adding, “An image becomes meaningful through its relationship with the story, performance, sound, and surrounding scenes.”
Vigneshkumar asserts that the creative process extends beyond generating visuals. He says, “I believe the art lies in the entire process, but the most important element is the creative vision of the person behind it. Sometimes the model gives me something beautiful, but it might not serve the story. I might reject an impressive-looking shot simply because it doesn’t communicate the emotion I want. That decision is also part of the art.”
The downside
These choices come with their own set of challenges. Filmmakers concur that maintaining continuity, in particular, proved difficult.
Shane explains that Gemini generates videos of about 10 seconds for each prompt, and these individual clips have to be edited together to create a feature-length film. “And in these 10-second videos, maintaining facial features, clothes, and even the hand placements and backgrounds proved tough,” he adds. Visual continuity is also another challenge. “The placement of objects, background details, lighting, and even the physical layout of a location can change unexpectedly between shots,” explains Vigneshkumar.
Muthu goes into detail about how he tackles the challenge: “Working with AI requires constant evaluation: whether a scene feels truthful, whether a character’s behaviour makes sense, and whether the film holds together… That requires patience, judgement, and a willingness to keep refining the work. I assess each scene by what it communicates, rather than its visual impact alone.”
While AI has lowered several production barriers, creators caution against equating access to technology with filmmaking expertise. Muthu points out, “AI can give more people an opportunity to begin, which is encouraging. However, time, resources, learning, and persistence still influence what someone can achieve.” He adds that storytelling, visual language, editing, and sound remain essential.
Vigneshkumar, who had to learn filmmaking fundamentals while experimenting with AI, opines the technology has made him more conscious of the craft. “AI has made filmmaking more accessible, but it hasn’t made the art of filmmaking effortless. The technology may reduce production barriers, but it still demands that we understand the craft. In fact, AI pushed me to learn a profession I once only dreamed about,” he says.
Though the debate over AI taking over jobs in the art industry continues to exist, Vigneshkumar says, “AI can generate what you ask for, but knowing what to ask for, what to keep, what to reject, and how to turn individual outputs into a meaningful story is where the filmmaker’s artistry comes in.”
Inside an AI director’s toolkit
Filmmaker Krishand RK, who is currently working on an AI-generated series in collaboration with an international studio, uses Luma AI and Kling AI for video generation, along with Midjourney for certain animation sequences. He points out that some use Meche for 3D renderings.
Vigneshkumar breaks down his toolkit:
Brainstorm & Screenplay Development: ChatGPT, Gemini, and Claude.
Visual Development: Nano Banana Pro, GPT Image 2, and Seedream 5.0, with Photoshop for corrections.
Video Generation: Google Omni 1.1, Seedance 2.5, and Kling, selected based on project requirements and budget.
“Every model has its own strengths and weaknesses, so I choose them based on what a particular scene or project demands,” he says.