At a time when artificial intelligence has begun redefining how information is created, distributed, and consumed, India’s media leaders have come together to confront one critical question: who controls trust in the AI era? At the ongoing AI Impact Summit 2026 in Delhi, industry voices have collectively acknowledged that while AI has unlocked scale and efficiency, it has also intensified the urgency around accountability, attribution, and public trust.
The panel discussion, “AI and Media: Opportunity, Responsibility, and the Road Ahead,” has examined how newsrooms, governments, and technology platforms have already begun navigating an ecosystem where information abundance risks diluting credibility. The session has repeatedly reinforced one central idea: AI may process information, but institutions have continued to produce trust.
The session brought together some of the most influential voices across Indian and global media, beginning with Sujata Gupta, Secretary General of the Digital News Publishers Association. She was joined by Kalli Purie, Vice Chairperson and Executive Editor-in-Chief of the India Today Group; Pawan Kumar, Deputy Managing Director at the Dainik Bhaskar Group; and Tanmay Maheshwari, Managing Director of the Amar Ujala Group. The discussion further featured Mohit Jain, COO and Executive Director at Bennett Coleman & Co; Navneet L V, CEO of The Hindu Group; and global industry representation from Robert Whitehead of the International News Media Association. The session was moderated by Ashish Pherwani, Partner – Media & Entertainment at Ernst & Young, who steered the conversation across policy, technology, and the future of AI-led journalism.
What followed was less panel discussion, and more manifesto.
“Press Doesn’t Produce Information. It Curates Trust.”
From the very first responses, one theme dominated: trust will be the last defensible moat in journalism.
Bennett Coleman & Co’s Jain captured the stakes bluntly: “Editorial discretion, verification and institutional memory are not optional. They are foundational. The press doesn’t just produce information, it curates trust.”
He warned that as AI commoditises information, trust will become scarce, and therefore valuable.
Hindu Group’s Navneet reinforced the same idea from an institutional lens: “Trust is not generated by technology. It is produced by institutions. That’s where you draw the line.”
In an era where synthetic content can mimic reality, credibility itself is becoming a premium product.
The ‘AI Sandwich’ Philosophy
Perhaps the most operationally clear framework came from India Today Group’s Purie, and it doubled as a philosophical position. “In AI, accountability must have a name.”
She described what she called an AI sandwich: Human intent starts the process. “AI assists in the middle and then human editorial judgment makes the final call.”
And she was blunt about the alternative: “You don’t want to become one biscuit in this AI cookie-cutter world.”
She also warned about “AI slop”, synthetic content that mimics credibility without accountability.
India’s AI Reality: Scale Without Uniform Literacy
Dainik Bhaskar Group’s Kumar shifted the conversation toward India’s unique consumption landscape: “We are not entertainment. We are news. If wrong information is consumed, it creates generational damage.”
He drew parallels to public infrastructure, arguing AI models and data ecosystems must be built with the same long-term national intent.
His warning was stark: “If news is not protected now, it will die in the next decade.”
AI As Tool, Not Replacement
Amar Ujala Group’s Maheshwari positioned AI as augmentation, not disruption: “We don’t see AI as replacement. We see AI as a tool to improve the quality of news.” His framing of AI architecture was particularly sharp, identifying content as the only variable layer across data, compute, and infrastructure stacks.
He also delivered one of the session’s most philosophical reminders: “We are just patrons of this world. We have to hand it over better.”
The Global Warning: AI Is Already Winning
International News Media Association’s Whitehead brought the international wake-up call: “You are already drinking AI water full of bacteria, and most people don’t know the difference.”
His biggest concern wasn’t future misuse, but current training on incorrect data and collapsing referral traffic economics for publishers.
The Nine-Point Charter: Media’s AI Survival Blueprint
At the heart of the session sat India Today Groups’ Purie’s nine-point framework, not as bullet-point policy, but as a systemic philosophy for AI-era journalism:
She argued the future must ensure fair value exchange for journalistic IP, calling for transparency in how AI systems ingest and process news content.
She pushed for mandatory attribution and traceability, not as optional ethical behavior but as embedded technological architecture, questioning why labeling AI output isn’t automated by default.
She positioned journalism as a public good, insisting algorithms must reward social impact, not just virality metrics.
She called for economic recognition of verified institutional content, warning that hallucinations cannot be treated as harmless quirks.
She highlighted the regulatory imbalance between legacy media and social platforms, demanding parity in accountability frameworks.
And perhaps most strikingly, she reframed attention itself as: “The rarest mineral we have.”
Her final principle demanded reciprocity from big tech, asking what platforms return in exchange for access to global attention economies.
Who Is Responsible When AI Goes Wrong?
Hindu Group’s Navneet cut through philosophical fog with operational clarity: “If legacy media is responsible for content, platforms must be held to the same standards.”
But he also admitted the industry has failed to communicate this urgency to consumers, arguing public demand for verified content must become bottom-up pressure.
The Sovereign Stack Question
On India-specific AI infrastructure, the panel was aligned: Foreign-trained models cannot fully represent India’s linguistic, cultural and socio-economic complexity.
Accuracy gaps in Indic language models were reportedly cited as low as ~50%, reinforcing urgency for domestic data ecosystems.
The Final Fear, And The Final Hope
Bennett Coleman & Co’s Jain perhaps delivered the most unexpected note: “If AI is anchored in trust, it might even take away loneliness. It could provide connection and comfort to millions.”
But the collective warning remained clear: If journalism funding collapses, AI will train on synthetic content, and reality itself becomes recursive.
As one line summed up the entire session: “If we don’t protect original reporting, AI will eventually just learn from AI.”
