Prasar Bharati goes shopping for AI talent to reinvent Doordarshan

NEW DELHI Public broadcasting and artificial intelligence are not phrases that usually sit comfortably in the same sentence, but Prasar Bharati is determined to change that. The broadcaster has put out a draft Expression of Interest for empanelling agencies, studios, start-ups and technology companies across nine categories spanning short-form content, digital humans, archive restoration and multilingual localisation, and it wants industry feedback by 21 September. This is not a tender. It is a call to arms for anyone with the chops to help drag Doordarshan and Akashvani’s vast archives, and its future output, into the AI era.

The scale of ambition is hard to miss. The EOI covers everything from thirty-second explainers to full-length feature films, AI anchors and synthetic newscasters, and the restoration of decades of Doordarshan and Akashvani footage using denoising, upscaling and colour enhancement. This sits atop WAVES OTT, the broadcaster’s streaming platform launched in November 2024, which is clearly the destination for much of this content. The pitch to industry is refreshingly candid: no minimum turnover requirement, capability trumps legacy credentials, and a pilot track exists specifically so promising start-ups without broadcast-scale portfolios can prove themselves through a paid proof-of-concept assignment.

What makes this genuinely interesting is the guardrails built around the opportunity. Digital humans and AI avatars, arguably the most reputationally fraught category for any state broadcaster, come with non-negotiable conditions: documented consent frameworks for any likeness based on a real person, visible on-screen disclosure whenever a presenter is AI-generated, and an outright ban on deepfakes of living public figures, elected officials or religious figures without explicit written approval. Provenance tracking under C2PA or equivalent standards is mandatory across the board, and applicants must disclose exactly which AI models they rely on, their country of origin and what happens if that model gets deprecated or locked down. This is not naive techno-optimism. It reads like a broadcaster that has thought hard about where AI content can go wrong.

The evaluation architecture backs up the intent. Applications are marked out of 100, with a 70-mark qualifying threshold and sectional minimums in technical, creative and delivery capability, evaluated by a committee expected to include external AI and digital media experts. Feature films and digital human deployments face extra scrutiny given the stakes involved. Crucially, empanelment itself confers no guarantee of work; commercial rates get discovered later through a limited QCBS-based RFP among empanelled agencies, refreshed roughly every six months, a sensible nod to how fast AI production costs are moving.

There is also a quiet nationalist undertone worth noting. Prasar Bharati has explicitly flagged a preference, though not a mandatory one, for agencies using Indian-developed AI, open-source ecosystems or India-hosted deployments, framing this as an AI sovereignty consideration alongside data residency compliance. For a broadcaster sitting on one of the country’s largest cultural archives, keeping the processing of that heritage within Indian jurisdiction is not an unreasonable instinct.

If executed well, this could be a genuinely useful template: a public institution using empanelment to build a diverse, technically vetted bench of AI partners rather than defaulting to a handful of large incumbents, while keeping ethics and provenance non-negotiable rather than an afterthought. The real test will come once actual assignments and rates start flowing. For now, though, Prasar Bharati deserves credit for asking the right questions before writing the cheques.