Broadcasters’ AI content rush outpaces India’s unsettled copyright protection rules

Indian broadcasters are building AI production capacity faster than the law is settling the question of who owns what comes out of it. Zee-backed micro-drama platform Bullet has rolled out Trinetra AI, an end-to-end filmmaking system that takes a story from script evaluation through AI-generated visuals to multilingual audio localisation. JioHotstar is expanding its own AI content slate as it prepares to enter the micro-drama format. Eros Innovation and Collective Artists Network have signalled similar plans. Each of these bets is being placed while the Copyright Act, 1957, the only statute that would determine whether any of this output can be legally owned, remains unclear about where AI-assisted work falls within it.

India has already tested this question, and reversed course
The clearest precedent broadcasters have to go on is not reassuring. In 2020, the Copyright Office granted registration to an AI-generated artwork called Suryast, listing the RAGHAV tool as co-author alongside its human creator, a global first at the time. The office subsequently withdrew that co-authorship recognition, reaffirming that the 1957 Act only allows natural persons to be named as authors. The episode suggests that fully AI-generated content, without substantial human creative input layered on top, may sit outside copyright protection altogether, precisely the category much of the new AI-assisted television and streaming output is drifting toward as production tools take on more of the script, visual and audio work themselves.

Why broadcasters, specifically, have more at stake than studios
Production houses that make a single AI-assisted title face a contained version of this problem. Broadcasters face a compounding one, because they are simultaneously creators and rights aggregators who license the same catalog across DTH, cable, OTT and overseas markets, often for years after first telecast. Every syndication and distribution deal typically requires a warranty of clean title to the underlying IP. A broadcaster that cannot clearly establish authorship over an AI-assisted show weakens the warranty it can offer buyers, complicates how that title is valued on a balance sheet or in an M&A process, and, most immediately, loses the ordinary legal tool, a copyright infringement claim, that it would otherwise use to pull down pirated re-uploads or stop a rival platform from lifting the same AI-generated format.

The risk runs in the other direction too
The gap not only threatens broadcasters’ ability to protect what they make; it can also expose them to liability for what their AI tools were built on. News agency ANI’s ongoing lawsuit against OpenAI, alleging its content was used to train ChatGPT without permission, is the clearest sign that Indian rights holders are already testing whether AI systems trained on unlicensed material infringe the rights of the original creators. Broadcasters have limited visibility into what footage, scripts or music the AI platforms they license, from in-house tools like Trinetra AI to third-party systems, were trained on. If any of that training data turns out to be unlicensed, the infringement exposure would attach to the broadcaster airing and monetising the output, not to the toolmaker whose training practices actually created the risk.

Reform is underway, but its timeline and shape are not
The government is not ignoring the gap. The Department for Promotion of Industry and Internal Trade published a working paper on generative AI and copyright in December 2025, and a separate expert panel formed the same year is reviewing the 1957 Act, with proposals reportedly including a formal definition of AI-generated works, clearer ownership rules and new accountability mechanisms for AI use. Neither process has produced a firm date for legislative change. That leaves broadcasters investing in AI catalogs now, ahead of rules that could later impose disclosure requirements or a stricter human-creativity threshold for protection, conditions that would apply retroactively to exactly the content being produced today under the current ambiguity.

That timing makes the current wave of AI investment a wager as much as a production strategy. The broadcasters moving fastest, Bullet’s Trinetra AI slate and JioHotstar’s expanding AI catalog among them, stand to gain the most from any eventual framework that recognises AI-assisted authorship, having built scale and audience data ahead of competitors. But they are also accumulating the largest libraries of content whose legal ownership rests on the least settled ground, with no clarity yet on whether protections finalised later will reach backward to cover what has already been made.