Walk into most Nepali newsrooms today and you won't find a robot writing the news. What you will find is AI quietly doing smaller jobs — translating a wire story, suggesting a headline, generating a quick illustrative sketch for a story that has no photo available. It's useful, unglamorous work. But as that quiet adoption grows, editors themselves are now asking a harder question in public: how far should this go before it starts working against the very things journalism is supposed to protect — accuracy, trust, and human accountability?
This piece looks at where AI is actually being used in Nepali journalism right now, why editors and press bodies are pushing for a formal ethical framework, the real risks involved, and how Nepal's approach compares with how other countries are handling the same challenge.
How Nepali Newsrooms Are Already Using AI
During a June 2026 workshop organised by Freedom Forum in partnership with Press Council Nepal, editors openly shared how AI tools have already found their way into daily newsroom routines. The most common uses reported were translation and language editing, headline selection, generating illustrative sketches for stories, and creating social media content to accompany published articles. Some newsrooms are also experimenting with AI-assisted fact-checking support, using tools to speed up verification of claims or cross-check details before publication.
At least one major Nepali digital news outlet, Nepal News, has gone further and published its own internal AI guidelines — reportedly the first Nepali media house to formally do so. Its policy requires that all AI-assisted content originate from a human-written draft, with AI used only to refine structure rather than generate facts from scratch, and mandates editorial review before anything reaches publication. Separately, Nepal's Online TV Journalist Association introduced its own AI code of conduct for online television journalism, reflecting a broader trend of individual media bodies moving faster than any single national regulation.
Editors' Calls for an Ethical Framework
The June 2026 Freedom Forum workshop — held in Kathmandu with support from the Ministry of Communication and Information Technology — brought together 54 participants, including six female editors, representing Press Council Nepal, the Department of Information and Broadcasting, and various Nepali media organisations and freelance journalists. A senior Press Council Nepal official told participants that AI can be a helpful tool but cannot replace human intelligence or the core values of journalism, while noting both the positive uses (news summaries, infrastructure sketches) and negative trends (fake photos and videos circulating on social media) that the council has observed.
By the end of the workshop, participants reached a joint commitment: Freedom Forum and Press Council Nepal agreed to prepare a draft national ethical and responsible AI use framework specifically for Nepali journalism, incorporating the suggestions raised during the discussion. This would mark one of the first sector-specific AI governance efforts in Nepal, distinct from the broader National AI Policy 2025, and specifically tailored to the unique pressures and responsibilities of a working newsroom.
The Risks: Misinformation and Job Displacement
Two concerns dominate the conversation among Nepali editors and researchers. The first is misinformation — editors have specifically flagged that AI often fails to fully grasp the nuances of the Nepali language when translating content, producing subtly inaccurate or misleading versions of a story if published without careful review. Combined with the broader wave of AI-generated fake photos and videos already circulating on Nepali social media (an issue that became especially visible during the March 2026 general election), there's a real risk that newsrooms could inadvertently amplify fabricated content if AI tools are trusted without independent verification.
The second concern is professional: as AI tools take over routine tasks like translation, summarisation, and basic content formatting, there are legitimate questions about what happens to entry-level journalism roles that traditionally built those exact skills. Academic researchers studying Nepal's media landscape have noted that this shift creates both opportunity and risk — AI can free up experienced journalists for deeper reporting and investigation, but only if newsrooms invest in retraining staff rather than simply cutting roles as tasks get automated. Copyright is a related friction point editors have raised directly: using AI tools trained on unclear data sources creates ambiguity around ownership and attribution that Nepali media law hasn't yet fully addressed.
What Responsible Use Looks Like in Practice
Based on what Nepali editors and early adopters have already put into practice, a few principles stand out as the working definition of "responsible" AI use in a newsroom context:
- Human drafts first, AI refinement second. Journalists write and verify the core reporting themselves; AI is used only to polish structure, language, or formatting afterward — never to originate facts.
- Mandatory editorial review before publication. No AI-assisted content goes live without a human editor checking it against the same standards applied to any other story.
- Clear internal guidelines, even before national rules exist. Nepal News and the Online TV Journalist Association both built their own AI policies rather than waiting for a mandatory national framework — a model other outlets can copy immediately.
- Prompt-writing and AI literacy training for journalists. Editors at the Freedom Forum workshop specifically called for capacity building on how to give AI tools appropriate prompts, since poor prompting is a major source of low-quality or inaccurate AI output.
- Transparent correction when AI-assisted content is wrong. Errors introduced through AI use should be corrected and disclosed the same way any other journalistic error would be, rather than quietly edited away.
Global Comparisons: How Other Countries Regulate AI in Media
Nepal isn't figuring this out in isolation — several other jurisdictions have moved further ahead on binding rules that Nepali policymakers and editors can learn from. The European Union's AI Act, for instance, introduces mandatory transparency obligations that take effect in August 2026, specifically requiring that AI-generated or AI-manipulated content published to inform the public on matters of public interest be clearly disclosed, unless it has undergone genuine human review with an identifiable editor taking responsibility for it. Deepfake images, audio, or video must be labelled as artificially generated or manipulated under the same rules, with only narrow exceptions for clearly satirical or artistic work.
Major international news organisations have also adopted their own internal standards well ahead of binding law — generally permitting AI for back-end tasks like transcription, translation, and headline suggestions, while prohibiting the publication of AI-generated images or video as if they were authentic news photography or footage, and requiring clear disclosure whenever AI substantially shaped a published piece. The common thread across these more mature frameworks is the same one Nepali editors arrived at independently: AI can assist journalism, but a named, accountable human editor must always remain responsible for what gets published.
Frequently Asked Questions
Is there an official AI policy for Nepali journalism?
Not yet as a binding national rule. Following a June 2026 workshop, Freedom Forum and Press Council Nepal committed to preparing a draft ethical framework specifically for AI use in Nepali journalism, though individual outlets like Nepal News have already published their own internal AI guidelines in the meantime.
Can Nepali journalists use AI to write news articles?
Current best practice, as reflected in early adopter policies, requires that journalists write and verify the core content themselves, using AI only to refine structure, language, or formatting — not to generate factual reporting from scratch without human verification.
What are the biggest risks of AI in Nepali newsrooms?
Editors have specifically flagged AI's difficulty accurately handling Nepali-language nuance in translation, the risk of inadvertently amplifying AI-generated misinformation circulating on social media, unresolved copyright questions around AI-trained content, and the potential displacement of entry-level journalism roles that traditionally taught core reporting skills.
Does Press Council Nepal monitor AI use in media?
Yes. A senior Press Council Nepal official has confirmed that the council regularly monitors AI use in Nepali media in accordance with the existing journalistic code of conduct, alongside its work developing a more formal AI-specific ethical framework.
How does Nepal's approach compare to global AI media regulation?
Nepal is still at the voluntary, framework-building stage, relying on individual newsroom policies and an in-progress ethical guideline, whereas jurisdictions like the European Union have already legislated binding disclosure requirements for AI-generated content published in the public interest, taking effect from August 2026.
Discussion