Nepal has drafted its first-ever National AI Policy, which was released for public consultation on February 4. The policy aims to regulate artificial intelligence technology, but experts argue it doesn’t fully address concerns related to data privacy and ethical risks.

Meanwhile, the Computer Association Nepal Federation (CAN Federation) has recommended stronger safeguards, clearer legal frameworks, and greater support for local AI innovation.

On Sunday, the federation presented its concept paper, ‘Use and Practice of AI in Nepal’, to Minister for Communication and Information Technology Prithvi Subba Gurung.

One major gap in the draft policy, the federation says, is the absence of a clear definition of AI tailored to Nepal’s specific needs. The report calls for an explicit definition that recognises AI as a system capable of autonomous data analysis and decision-making.

“A clear, precise definition will distinguish policy formulation from general discourse,” said Chiranjibi Adhikari, CAN’s vice president and AI policy coordinator.

The document also suggests including a glossary of technical terms such as machine learning, data privacy and AI governance for a more informed policy-making process.

In addition to clarifying terminology, the CAN Federation advocates the need for AI integration across key sectors including healthcare, agriculture, and disaster risk management.

This includes AI-driven telemedicine services for rural populations, precision farming technologies to enhance crop yields, and AI-based weather forecasting models to help mitigate disaster risks.

“Data security remains a pressing issue,” Adhikari says. “The use of AI must not come at the cost of privacy.”

For the security concerns, the expert committee which prepared the concept paper calls for the implementation of stringent AI data governance protocols. The committee includes Ganga Bhandari, IT specialist at UNDP Nepal; Prof Dr Sudan Jha, AI expert at Kathmandu University; Advocate Dikchya Raut, tech and policy analyst; and vice president Adhikari.

According to the committee, these protocols should focus on data safety, confidentiality, and security, in alignment with the core principles of the CIA triad—confidentiality, integrity, and availability.

“A separate section on AI data governance is needed, covering aspects such as data cleansing, data safety, data secrecy, and security,” Jha said. “Nepal also needs to establish its own AI data centres to reduce dependence on foreign entities.”

Cross-border data-sharing policies is yet another necessity that the concept paper points out. “We need to balance local needs with international standards while drafting AI policies,” Jha added.

For a sustainable AI ecosystem, the federation stresses investment in human resources. “In the current digital era, Nepal requires structural initiatives for the sustainable development of AI technology,” said federation president Sunaina Ghimire Pandey. “AI has great potential in education, healthcare, agriculture, infrastructure, and administration.”

Experts have raised concerns about the lack of Nepal-specific provisions in the draft policy. They recommend AI literacy programmes in schools, scholarships for AI research, and the creation of dedicated AI research centres.

The report also recommends AI technologies tailored to Nepal’s linguistic needs with a focus on Nepali and regional language models.

Another point raised in the recommendation is the assurance of ethical and inclusive AI. The federation emphasises AI regulatory framework that prioritises ethical considerations and protects marginalised communities from potential biases in AI applications.

“Our aim is to implement the AI policy in a transparent and inclusive manner,” said Shaligram Parajuli, an IT expert from the Ministry of Communication and Information Technology. “We will develop a clear roadmap based on the input, concerns, and support of all stakeholders and will soon finalise the policy and implement it.”