When the AI-made series Mahabharat: Ek Dharmayudh was released, a single scene from episode one went viral.
A plush, modern bedside table straight out of Ikea lies in the corner of the frame, next to a period correct bed. The comments, like this one: "just missing a wireless charger", were brutal. I saw its first two episodes on Jio Hotstar, and this bizarre anachronism is at the heart of India's first grand, "AI-powered" mythological series. A project that had grand ambitions but with an execution so bad it seems to have forgotten the very soul of the story it was allegedly trying to tell.
The PR by Collective Artists Network for this series that released on October 25 on JioHotstar, touted it as a revolution, "when a timeless epic meets groundbreaking technology", "a new benchmark for Indian entertainment" designed to "build a bridge between tradition and the future", that AI would help "breathe life into ancient legends with breathtaking scale and emotional depth". What emerged, far from meeting its lofty ambition, has become a stark, witty, and painful lesson in the perils of prioritising technological novelty over basic filmmaking craft.
Without intending to, this newest, worst avatar of one of the world's oldest, greatest epics has become an inadvertent masterclass in how not to approach filmmaking with the latest AI tool.
The soulless spectacle: Let’s begin with the most evident flaws. The frames are often lush with colour and detail, an advertisement for AI's ability to generate grand, eye-catching vistas. Yet it quickly turns stale when the characters open their mouths to speak.
Their faces often slip into an uncanny valley with awkward mouth movements and expressions failing miserably to match the emotional weight of the dialogue. The result is that key, emotional moments fall flat. This is understandable, considering that the first few episodes might have been made when the tech used to create them was still raw. But why use it in the first place?
The series' cardinal sin is its hollowness. The Mahabharata is not a sequence of events but a play, a cosmic war of duty, morality, and chaos, driven by characters with psychological complexities so immense that it is said that "whatever is here, may be found elsewhere; but what is not here, is nowhere". To reduce this to what a reviewer scathingly called "a cheap AI-generated fever dream" is a profound error in judgment to punish which Lord Krishna could consider employing the Sudarshan chakra.
Because when the good lord delivers the Gitā's profound wisdom, like my favourite – you only have the right to your duty, not its fruits, he shouldn't sound like a "motivational YouTuber" or a "crypto bro". Yet, look at the first two episodes, and you see how the epic's divine pathos are drained, leaving behind a beautifully rendered but emotionally vacant puppet show.
A masterclass in how not to use AI: For all its flaws, the show is not without merit, but it is one its makers never sought. To me, it serves as an invaluable case study of critical pitfalls to avoid. It also underlies what I think is a fundamental truth: AI does not replace the need for filmmaking craft; it amplifies it. To understand this, let us take a few examples from episode one alone.
At about 10:42, the arrow in the quiver of King Shantanu lacks the tails it had seconds prior. Then, he is seen holding an arrow, followed by a random cut to an arrow stuck in a tree, then back to the king still holding his arrow with no context in the dialogue or the scene. These are not AI failures; they are failures of basic directorial supervision and editing. AI generates assets, but it is the human filmmaker's job to ensure narrative and visual consistency.
Then, throughout Episode one, King Shantanu wears full armour, even when sitting in a palace, talking to his ministers. Why? Does he think one of them – dressed in regal splendour – will come and try to attack him?
I know why this happened: perhaps this was the only way to maintain character consistency in the early days of Midjourney, Luma AI, or Kling. But ways could still have been found around it.
Then there are the wide shots of the palaces, which look no different than colourless ruins – except that they are complete ruins. The makers didn't even have the wisdom to realise that how places look today is not how they were in their glory days. Hence, without a skilled production designer and art director guiding the AI with a coherent vision, the world feels random and disjointed. The limitations of AI tools in filmmaking are not an excuse for a lack of imagination; it demands more creativity to work within or cleverly disguise those limits.
But, the biggest problem with the story is the most human of all: the screenplay.
Mahabharata is chock-full of dramatic stories, moral dilemmas, and turns of events. In the hands of an able screenwriter, it'll lend itself beautifully to hooks and cliffhangers. That Ek Dharmayudh does not talks about the deficiency of the humans at the helm, not the machine.
AI can be a phenomenal assistant scriptwriter, a tool for brainstorming concepts and research, and a trigger for original ideas in creators. Yet, here, the writing is so atrocious that it is a dire disservice to one of the greatest stories ever told. The focus was on using AI for the visuals, while the narrative's foundational pillar was neglected entirely, leading to content that is neither strong in its AI nor in its writing.
From creative artists to creative "media": The backdrop to this series also adds a layer of irony to its miscalculations. The project is credited to the Collective Media Network. Google that name, and you won't find a company with that name associated with content creation. Hence, this appears to be a hastily made subsidy, perhaps to divert attention from their core competence as the Collective Artists Network (CAN), an agency representing some of India's most brilliant creative minds. This might also have been after they faced backlash from people like Anurag Kashyap on the use of AI. Replacing 'Artists' with 'Media', seems like a delicious metaphor of the AI age.
This seems paradoxical: in seeking to deploy a new technology, they are seemingly distancing themselves from the very community – the artists – who could have infused the project with the soul it so desperately lacks. The result is a project that does a double disservice: it undermines the talent CAN represents, while also setting an extremely low, slop-like benchmark that will unfairly colour the perception of emerging, serious AI filmmakers in India. When a genuinely thoughtful AI-assisted film emerges, audiences, already turned off by this experience, may be unwilling to give it a chance it might deserve.
Don't race to be the first, strive to be the best: Mahabharat: Ek Dharmayudh is not bad because it used AI; it is unwatchable because it used AI poorly. It confused the ability to generate images with the artistry of storytelling. In prioritising the headline-grabbing goal of being "India’s first AI-powered mythological series" over a more meaningful ambition of being India's best-told one, it does a disservice to both the story and the tools they used to make it.
The lesson for filmmakers from this series is clear: AI is a tool that demands more talent, not less. It requires a new kind of professional: one who is not only well-versed in various aspects of filmmaking but also in history, art, and the nuances of guiding an AI to serve a coherent, creative vision of the filmmakers. AI should always be treated as a collaborator, not a replacement.
Mahabharata has survived many a millennium, through countless retellings for one simple reason: each version was infused with human imagination, emotion, relevance, and reverence. This version, in its rush to be in tune with the future, becomes instead a dire reminder that in the jugalbandi between tradition and technology, what should always lead is the soul.
from Hindi https://ift.tt/f36OLQn
via IFTTT
