Human oversight cannot be outsourced to AI models, says MIB’s Prabhat

While acknowledging AI’s transformative potential, he stressed that newsroom accountability cannot be automated

author-image
Lalit Kumar
New Update
Prabhat
Listen to this article
0.75x1x1.5x
00:00/ 00:00

New Delhi: As artificial intelligence tools become embedded across Indian newsrooms, human editorial oversight must remain non-negotiable, said Prabhat, Additional Secretary, Ministry of Information and Broadcasting (MIB).

“Human oversight as a non-negotiable editorial responsibility cannot be outsourced to a model,” Prabhat said, delivering one of the strongest cautionary notes at a Google News Initiative event.

While acknowledging AI’s transformative potential, he stressed that newsroom accountability cannot be automated.

“AI can assist but it cannot approve the final thing,” he said. Prabhat described the current AI wave as a defining opportunity for India and the Global South. “The AI moment offers an unprecedented opportunity to the global South,” he said, adding that India is positioned to lead the transition.

He highlighted multiple use cases where AI is already enhancing newsroom efficiency, from summarising long reports and analysing metadata to translation, transcription and deepfake detection.

Press Trust of India, for instance, is using AI-powered tools to generate infographics 12 times faster than before, while Akashvani is translating “Mann Ki Baat” into 24 languages in real time using AI.

However, speed and cost reduction, he warned, cannot come at the cost of credibility. “The competitive advantage in the AI age will not be speed, it will be credibility,” Prabhat said. He proposed building what he termed a “trust architecture” for AI in journalism, a structured framework combining disclosure, verification protocols, clear accountability lines and data discipline.

“Disclose where it matters,” he said, urging newsrooms to inform audiences when AI has materially shaped content.

AI drafts, he advised, should be treated as “unverified working papers,” not publishable material, and must undergo editorial scrutiny. Prabhat also emphasised privacy and source protection. AI convenience, he cautioned, must not become a pathway to compromising sensitive journalistic material.

From a policy standpoint, he reiterated that the government aims to enable innovation while safeguarding the information ecosystem. “The government is committed to enable innovation, protecting the information ecosystem,” he said.

In a pointed message to technology platforms and media houses, he urged stakeholders to move beyond technological enthusiasm. “Move from tools to trust,” Prabhat said. He concluded by calling for practical, implementable standards that ensure AI strengthens journalism rather than destabilising it.

Ministry of Information and Broadcasting MIB AI in newsrooms
Advertisment