Over 60 countries around the world have already published national strategies for Artificial Intelligence. The United States, the European Union, China, Singapore and even our neighbouring India and Bangladesh have all formally defined how AI will be developed, regulated and used within their borders. Nepal, for most of the past decade, was not among them, and that absence carried a cost. In August 2025, Nepal finally approved its National AI Policy 2082, a significant milestone. But having a policy on paper and making it work on the ground are two entirely different things.
For readers who are not familiar with the term, Artificial Intelligence can be described as computer systems that can perform functions that would otherwise require human intelligence, such as language and image recognition, decision-making, or finding patterns in large volumes of data. When you talk to a voice-recognition program on your phone, when your bank tells you a transaction is suspicious, or even when the software in a hospital scan identifies cancer, that is AI at work. It is not science fiction. Millions of Nepalis have already been exposed to it and many of them are unaware that they engage with its services on a daily basis.
Governments around the world realised early that AI was not to be solely left in the hands of the private sector. The EU adopted the first global AI legislation, the EU AI Act, which became effective in August 2024. It categorises AI systems according to the level of risk: those that present unacceptable risks, like social scoring by governments, are prohibited completely. The high-risk systems, such as AI in hiring or medical diagnosis, have to be transparent and safe to work. The EU’s approach is clear: technological progress must not come at the cost of human rights or public safety.
Closer to home, India unveiled its IndiaAI Mission in March 2024, which was supported by a budget of more than Rs 10,371 crore or about NPR 166 billion over a period of five years. The mission is built around seven key pillars, which include computing infrastructure, AI innovation, data access, AI applications, workforce training, startup financing and ethical AI. With subsidised rates of as little as Rs 65 per hour, India is constructing more than 18,000 specialised computing units that are available to researchers and startups. It has established AI Data Labs nationwide, is investing in 13,500 AI scholars, undergraduate through PhD and is creating AI models that are trained on Indian languages and local conditions.
The National AI Policy 2082 adopted by Nepal, which follows months of inter-ministerial consultation, is a true step in the right direction. The policy provides a legal and regulatory framework to govern AI, prioritises ethical and transparent applications of AI in all areas and includes the prospective establishment of a National AI Centre and an AI Regulation Council. It has a goal of training at least 5,000 qualified AI specialists, incorporating AI programs at schools through universities and fostering government-business connections. These are precisely the types of foundations that a developing country should have on paper.
What is of concern, though, is not the ambition of the policy but its implementation. A well-known international index that ranks AI readiness in 193 countries places Nepal at position 150, which indicates acute deficits in infrastructure, highly skilled labour, and the ability to regulate. The policy does not stipulate a particular budget. It states about the public-private collaboration but does not provide any tangible incentives to invest. A policy that seems good in a paper but has hit a snag in execution is of no use to anybody.
The possibilities are valid and worth struggling with. Artificial intelligence may transform the healthcare provision in Nepal’s remote mountain and hill districts, where physicians are rare and diagnostic equipment is absent. It might enhance the agricultural prediction for farmers who remain reliant on climatic patterns that are becoming more unpredictable with climate change. In education, the AI-based individualised learning mechanisms may also be used to access students in villages where the problem of teacher shortage is endemic. In the field of public administration, AI has the potential to decrease bureaucratic delays and enhance service delivery, which, in Nepal, is severely lacking. These are not far-fetched possibilities. These gains have been proven by countries with similar economic positions.
To fulfil its promise, several things need to occur immediately in order to make the National AI Policy 2082 effective. One, a specific budget has to be dedicated to it – a policy that is not funded is a press release. Second, the AI Regulation Council should be formed as soon as possible and manned by truly independent experts, rather than political nominees. Third, data governance should not be overlooked: Nepal does not have appropriate structures of data quality, data security, and privacy of citizens and AI created on bad or uncontrolled data will yield bad and even detrimental results. Fourth, the issue of brain drain should be addressed directly. Nepal produces good engineers who go to seek better opportunities elsewhere. The talent gap will not disappear without the policy providing competitive opportunities domestically, in the form of research grants, incubation of startups and rational public sector AI positions.
There is also the question of public awareness. The AI policy cannot be a discussion between the ministries of the government and technology experts only. The common people in Nepal should be made aware of what AI is, what their rights are in case AI systems make decisions about them and what they can do when there are errors. Countries that have succeeded with AI policy have invested in public engagement alongside technical infrastructure and Nepal must do the same.
Nepal stands at a genuine turning point. The National AI Policy 2082 is an indication that the government is aware of the stakes. The question now would be, will that signal be accompanied by resources, institutions and accountability? The world is no longer waiting. India is incurring billions. The EU is a law-bound body. China is establishing world standards. In the case of a small nation like Nepal, the time to influence the development of AI within the country, instead of it being imported as a byproduct of external mechanisms with external regulations, is limited. The starting point is the policy. Now is when the actual work begins.
