Illustration: Katmandu Journal
A MEDIA market is not dying because of artificial intelligence. It is dying because it never truly lived. This is the reality in Nepal, where a decade after newspapers staggered onto the internet and radio hosts became TikTok commentators, the industry still rests on three brittle legs: government tenders, telecoms sponsorships and banner ads from banks and cement companies with questionable taste in fonts. Now, as AI upends global advertising into an automated, data-driven science, those legs are buckling fast. The apocalypse has arrived early, and there may not be enough revenue to print a funeral notice.
Advertising the world over is undergoing a coup. AI can generate polished video content, personalise ad targeting, optimise campaign performance and do all of this for a fraction of the cost of a human-led agency. The impact is most visible in the West, where creative firms are retooling entire workflows. But in places like Nepal the consequences are more existential. With scant legacy infrastructure and a market historically addicted to Facebook boosts and Google crumbs, Nepali media may not get the chance to adapt. It may simply be leapfrogged into obsolescence.
The tools driving the shift are not theoretical. Meta’s Advantage+ campaigns can now generate creative assets and manage ad delivery automatically. TikTok’s Creative Studio churns out video reels in seconds. Even entry-level businesses—momo stalls, tuition centres, home-baking startups—are discovering that a 3,000-rupee AI-generated video can outperform a professionally shot television advertisement. For advertisers the appeal is obvious. For local ad agencies, many of which still traffic in jingle-based nostalgia, it is an extinction-level event.
The ramifications for the news industry are even grimmer. For years publishers relied on search-engine-optimised “explainer” articles to drive traffic: how to apply for the American visa lottery, the best phones under 30,000 rupees, who married whom. That model now faces annihilation. Google’s new AI Overviews answer user queries directly on the search page, bypassing publishers entirely. Business Insider, an American outlet, lost 55% of its search traffic over three years. Some publishers are preparing for “zero” Google traffic. Nepali outlets, operating on shoestring budgets and content mills, have even less room to manoeuvre.
The wider risk is that journalism itself becomes invisible. AI models train on publicly available content, extract the facts and spit out answers, often without credit. If a Nepali newsroom produces an investigation into hospital corruption, it may inform an AI chatbot’s response to a user query, but generate no clicks, no ad impressions, no revenue. Without structural data tags that help machines identify sources, even the bots may struggle to attribute the story. Content becomes ambient: read by machines, acknowledged by none.
To make matters worse, AI-driven advertising is increasingly platform-native. Google and Meta deliver ads but they also plan, create and iterate them. That leaves little reason for advertisers to go through intermediaries, let alone fund journalism. A Kathmandu wedding photographer can now target brides in Pokhara more precisely using TikTok’s algorithm than via a full-page ad in a popular online news portal. The fallout is the disintermediation of the media. Journalism no longer sits between brands and audiences. Algorithms do.
Some in the media scene cling to the hope that prestige or tradition will insulate them. But the algorithm is ruthlessly indifferent to institutional legacy. It rewards engagement rather than editorial standards. Even now many Nepali newsrooms confuse Facebook reach with actual influence or mistake YouTube subscriptions for sustainability. Few invest in user analytics or reader journeys. Fewer still ask what their audience might pay for, let alone trust. The gap between journalistic production and audience consumption is growing, and AI threatens to snap it altogether.
That disjuncture opens the door for public relations to masquerade as news. Politicians, brands and bureaucrats may soon publish AI-optimised articles designed not for voters or consumers, but for language models and search engines. In such a world influence flows not through column inches but through prompt engineering. The future media strategist is not a journalist, but a coder.
There is, however, a sliver of opportunity. As the attention economy fractures and commodified content floods every feed, trust may regain its premium. Publishers who cultivate direct relationships—through newsletters, memberships, events—may find lifelines. Journalism that is human, distinctive and deeply reported could thrive amidst the sludge. But this means abandoning fantasies of passive growth. No amount of search optimisation can save a newsroom nobody remembers.
The media sector, guilty of borrowing digital strategies from the West half a decade too late, now faces a chance to leapfrog. It can build newsrooms around loyalty, not virality. That means understanding that clicks are cheap, but convictions are costly and worth more. It means producing fewer stories, better packaged, with clearer purpose. And it means imagining a future where journalism is a service and not a commodity.
That future is neither inevitable nor guaranteed. Most Nepali outlets are unprepared. Budgets are shrinking. Newsroom morale is waning. And many editors still believe engagement can be gamed. But survival is no longer a matter of tweaking headlines or buying traffic. It is a matter of being indispensable, to readers.
AI may well be the final straw for Nepal’s precarious media. But it also offers a clarifying moment. The ad economy it upended was never particularly just or sustainable. What replaces it could be better, if anyone remains standing to build it. ■
