The Nepalese government’s recent release of a draft Artificial Intelligence (AI) policy has sparked widespread debate among stakeholders, raising concerns about its feasibility, inclusivity, and ability to address critical issues like data protection, algorithmic bias, and cybersecurity. While the policy aims to regulate AI development and usage, experts argue that it lacks clarity in implementation strategies, prioritization, and resource allocation, potentially undermining its effectiveness.

The Ministry of Information and Communication invited public feedback on the draft policy, prompting discussions among AI experts, civil society organizations, and advocacy groups. However, critics contend that the drafting process excluded key stakeholders, including grassroots communities, labor groups, and human rights advocates. This exclusion has fueled skepticism about whether the policy can adequately address the needs of marginalized populations or mitigate potential risks associated with AI technologies.

Lack of Prioritization and Actionable Plans

One of the primary criticisms leveled against the draft policy is its failure to prioritize specific areas for immediate action. While the document touches upon various aspects of AI, such as governance, ethics, and innovation, it does not specify which issues require urgent attention. Digital Rights Nepal and the AI Association of Nepal have jointly highlighted this gap, arguing that without clear prioritization, implementing the policy could become chaotic and ineffective.

Santosh Sigdel, Executive Director of Digital Rights Nepal, emphasized the absence of a robust enforcement mechanism. “The policy lacks actionable plans, timelines, and resource allocation,” he stated. “Without these elements, it risks remaining theoretical rather than practical.” Sigdel further noted that the absence of a clear roadmap could hinder efforts to address real-world challenges, such as misuse of personal data, biased algorithms, and cybersecurity threats.

Data Protection Concerns

A significant concern surrounding the draft policy is Nepal’s inadequate legal framework for data protection. Currently, the country lacks comprehensive laws addressing data security and privacy, leaving individuals and institutions vulnerable to misuse of sensitive information. Although the draft mentions misinformation, critics argue it fails to outline sufficient technical or legal measures to ensure data integrity.

Stakeholders have called for stronger mechanisms to protect personal data, particularly in sectors like health, finance, and government services. Shailendra Raj Giri, President of the AI Association of Nepal, stressed the need for advanced encryption methods, access controls, and breach prevention protocols. He also advocated for clear directives on data usage, regular monitoring mechanisms, and legal recourse in cases of violations.

The current legal frameworks, such as the ‘Electronic Transactions Act, 2063’ and the ‘Cybersecurity Policy, 2080,’ are deemed insufficient to handle the complexities posed by AI systems. Experts recommend establishing robust legal and institutional structures for data ownership, sharing protocols, and protection against algorithmic bias.

Calls for Ethical Oversight

To ensure safe and equitable AI deployment, stakeholders propose forming a dedicated AI Ethics Committee. This body would oversee transparency, accountability, and adherence to ethical principles. Critics argue that the proposed AI Regulatory Council (Section 10.1), dominated by ministry representatives, focuses more on regulation than governance. They advocate for an independent committee that includes diverse stakeholders, such as academics, technologists, and civil society members.

Giri highlighted the importance of developing grievance redressal mechanisms and ensuring public participation in shaping AI policies. “An AI Ethics Committee should evaluate risks posed by automated decision-making systems and safeguard citizens’ rights,” he stated. Such a committee would play a crucial role in mitigating discriminatory practices, ensuring fairness, and fostering trust in AI technologies.

Human Oversight and Accountability

Critics stress the necessity of human oversight in AI-driven decisions. They demand mandatory human review systems and appeal mechanisms for individuals affected by AI outcomes. Automated processes must include provisions for periodic audits and corrections to prevent harm. Stakeholders also call for transparency in how AI systems arrive at decisions, enabling users to understand and challenge outcomes when necessary.

Sigdel emphasized the need for inclusive consultations during the policy formulation process. “Marginalized communities, labor groups, and human rights advocates must have a voice in shaping AI policies,” he stated. Without their input, the policy risks overlooking critical issues like gender disparities, socioeconomic inequalities, and the impact of AI on vulnerable populations.

Cybersecurity Challenges

As AI systems increasingly rely on large-scale data processing, they become attractive targets for cyberattacks. Stakeholders warn that Nepal’s AI policy must prioritize cybersecurity measures, secure database management, and automated risk assessment tools. Current safeguards are insufficient to protect AI systems from malicious actors seeking to exploit vulnerabilities.

Experts urge the government to adopt proactive strategies, such as implementing end-to-end encryption, strengthening access controls, and conducting regular vulnerability assessments. Additionally, they recommend establishing incident response teams to address breaches promptly and minimize damage.

Addressing Algorithmic Bias

Algorithmic bias remains a pressing concern, as AI systems often reflect the prejudices and limitations of their creators. Critics argue that the draft policy does not adequately address how to identify and mitigate biases in AI algorithms. They call for transparent methodologies to evaluate datasets and ensure fairness in decision-making processes.

Giri proposed incorporating gender perspectives and analyzing AI’s impact on marginalized communities. “We need to assess how AI technologies affect different social groups and take corrective actions to promote equity,” he stated. By addressing these issues, Nepal can harness AI’s potential while minimizing its adverse effects.

Implementation Challenges

Despite its ambitious goals, the draft policy faces significant implementation challenges. Critics point out that Nepal lacks the technical expertise, infrastructure, and financial resources required to execute the proposed initiatives effectively. Without addressing these gaps, the policy risks falling short of its objectives.

Stakeholders recommend allocating adequate funding, building local capacity, and fostering international collaborations to overcome these barriers. They also emphasize the need for continuous monitoring and evaluation to ensure compliance with ethical standards and legal requirements.