In August 2025, Nepal's Cabinet approved the National AI Policy 2082, and with it came something the country had never had before: a dedicated institutional structure for overseeing artificial intelligence. At the centre of that structure sits a body that most people now search for as the AI Regulatory Authority Nepal — though its formal name in the policy text is the AI Regulation Council. Some commentators and early drafts also referred to it as the AI Supervision Council. In this guide, we unpack who actually sits on this body, what powers it holds, how it reports upward, and where it fits alongside the separately created National AI Centre.
Why Nepal Needed a Dedicated AI Body
Before 2025, Nepal's digital rulebook was scattered across the Electronic Transactions Act, the ICT Policy, the Digital Nepal Framework, and the National Cybersecurity Policy — none of which were written with generative AI, algorithmic decision-making, or automated credit scoring in mind. As AI tools moved from experimental to mainstream across banking, healthcare, and public service delivery, policymakers recognised that a single coordinating body was needed to set standards, monitor compliance, and keep pace with a fast-moving technology. That recognition is what produced the AI Regulation Council.
Who Sits on the Council — and Who It Reports To
The AI Regulation Council is chaired by the Minister for Communications and Information Technology, which places it firmly under the umbrella of the Ministry of Communication and Information Technology (MoCIT). Its proposed composition, as outlined during the policy's drafting stage, brings together an unusually wide mix of regulators and administrators, including:
- The Governor of Nepal Rastra Bank (the central bank)
- Secretaries from the Ministries of Communication, Finance, Industry, and Education, Science and Technology
- The Secretary of the National Planning Commission
- The Chief Executive Officer of the E-Governance Board
- The Chairman of the Nepal Telecommunications Authority (NTA)
This composition matters because it signals that AI regulation in Nepal is not being treated as a purely technical or telecom issue — the presence of the central bank governor alone tells you that financial-sector AI (credit scoring, fraud detection, digital payments) sits high on the priority list. Certain appointed members serve three-year terms, and the Council is required to meet at least twice a year, with the option to invite additional officials or subject experts as needed. Ultimately, the Council answers to the Cabinet through MoCIT, while broader monitoring of how well the policy is being implemented also falls to the Office of the Prime Minister and Council of Ministers and the National Planning Commission.
Its Mandate: Compliance, Transparency, and Standards
The Council's job is deliberately broad rather than narrowly technical. According to the policy framework, its core responsibilities include providing guidance for effective implementation of the AI policy, setting quality and performance standards for AI-related products and services, and conducting awareness and capacity-building programmes across ministries and the public. In practice, this makes the AI Supervision Council Nepal function less like a licensing regulator (in the way a securities board might license brokers) and more like a policy-steering and standard-setting body that shapes how other agencies — banks, telecom operators, hospitals, and public offices — are expected to govern their own AI use.
Transparency and ethical use run through nearly every function assigned to the Council: reviewing whether AI systems respect data privacy and citizens' rights, ensuring non-discriminatory outcomes, and pushing sector regulators to update outdated laws so they remain relevant to AI-driven decision-making. It is this transparency mandate that connects the Council to a wider risk-classification exercise, which we cover in detail in our companion article on Nepal's AI risk classification framework.
How It Differs From the National AI Centre
This is where a lot of confusion sets in, so it is worth being precise. The AI Regulation Council and the National AI Centre are two distinct bodies with two distinct jobs:
- AI Regulation Council — the high-level, policy-setting body made up of secretaries, the central bank governor, and senior regulators. It meets periodically, sets direction, and approves standards. Think of it as the board of directors.
- National AI Centre — the operational, day-to-day implementing body established under MoCIT. It serves as the Council's secretariat, conducts research, coordinates the provincial AI Excellence Centres, maintains the National AI Index, and carries out the practical work of monitoring AI systems and preparing risk-mitigation action plans.
In short, the Council decides what the rules should say; the National AI Centre is the machinery that carries those rules out on the ground. Companies and developers dealing with compliance questions will, in most cases, interact directly with the National AI Centre rather than the Council itself.
Early Actions Since Formation
Since the policy's approval, the most concrete regulatory move has actually come from a Council member rather than the Council as a whole: Nepal Rastra Bank drafted a dedicated AI guideline for banks, finance companies, and digital payment service providers, covering everything from credit scoring and fraud detection to board-level accountability for AI-driven decisions. This is widely seen as an early signal of how the broader AI Regulation Council framework is expected to filter down into sector-specific rules. Beyond the financial sector, early-stage work has focused on setting up the National AI Centre itself, scoping the planned regulatory sandbox, and beginning consultations on AI Excellence Centres at provincial universities.
Criticism and Open Questions
Not everyone is convinced the structure will work as intended. Industry commentators have pointed out a few recurring concerns:
- Implementation over paperwork: Nepal has a track record of publishing ambitious digital policies — the Digital Nepal Framework being one example — that moved slowly once it came to actual execution. Sceptics ask whether the Council will have the budget and staffing to match its mandate.
- Overlapping authority: With the central bank, the telecom regulator, and multiple ministries all sitting on one Council, there is a real question about how disputes over jurisdiction — for example, who regulates an AI-powered fintech app — will be resolved in practice.
- Private-sector involvement: Several industry voices have argued that without meaningful participation from AI companies and startups (not just government secretaries), the Council risks producing standards that are disconnected from how AI is actually built and deployed.
- Naming confusion: The interchangeable use of "AI Regulation Council," "AI Regulatory Authority," and "AI Supervision Council" across government drafts and media coverage has made it harder for businesses to know exactly which body to approach for guidance.
Frequently Asked Questions
Is the "AI Regulatory Authority" an official government name?
Not exactly. The National AI Policy 2082 formally establishes the "AI Regulation Council." However, because its function is regulatory and supervisory in nature, media outlets and industry commentary frequently refer to it as the AI Regulatory Authority or AI Supervision Council. All three terms currently point to the same body.
Does the Council issue licenses to AI companies?
Based on the current policy text, the Council's role is to set standards, provide implementation guidance, and oversee compliance — not to run a licensing regime in the way a securities or telecom regulator might. Sector regulators, such as Nepal Rastra Bank for financial institutions, are expected to translate these standards into binding, sector-specific rules.
Who should a fintech startup contact for AI compliance questions in Nepal?
In most practical cases, the National AI Centre — as the operational secretariat — is the first point of contact, alongside the relevant sector regulator (for example, Nepal Rastra Bank for banking and payments-related AI use cases).
How often does the Council meet?
The policy requires the AI Regulation Council to meet at least twice a year, though it may convene additional sessions and invite outside officials or experts as required.
Discussion