Bangladesh is moving to put AI under a firm regulatory lens, with the interim government unveiling a draft National Artificial Intelligence (AI) Policy for 2026-2030 that significantly expands state oversight of data, algorithms and high-risk applications.
Released on Wednesday for public consultation, the draft signals a cautious, security-first approach to governing rapidly advancing technologies.
At the same time, it seeks to balance control with innovation, proposing new funding mechanisms, regulatory sandboxes and skills development initiatives to ensure the country can adopt AI responsibly without falling behind globally.
The draft emphasises digital sovereignty and the protection of national data assets, proposing reinforced controls over the collection, sharing and use of data within AI systems.
While it allows secure data exchange among government agencies under the National Data Management Ordinance 2025, the framework prioritises state oversight of sensitive and strategic datasets, placing limits on how such data may be accessed and reused.
Significant attention is given to AI-generated content, particularly deepfakes, online rumours and misinformation.
The policy outlines measures to prevent the creation and dissemination of AI-generated false or misleading information, aligning enforcement with the Cyber Security Ordinance 2025.
These provisions extend AI governance into areas already covered by existing cyber and digital security laws.
The draft adopts a risk-based regulatory structure under which certain AI uses are explicitly restricted or prohibited. Practices such as social scoring systems and mass surveillance are banned outright.
At the same time, it allows regulated use of high-risk AI systems by public authorities-including in law enforcement, healthcare, employment and public service delivery-subject to safeguards and mandatory impact assessments.
Human rights, transparency and accountability are identified as non-negotiable principles of the policy.
For high-risk AI systems, the draft mandates algorithmic impact assessments and introduces safeguards intended to prevent discrimination and harm.
These requirements apply particularly to systems that may affect citizens' rights, access to services or legal outcomes.
The policy builds on earlier legal frameworks introduced by the interim government, including the Personal Data Protection Ordinance 2025.
Together, these instruments establish the legal infrastructure for AI governance, linking data protection, cybersecurity and AI regulation within a unified, state-led framework.
Beyond regulation, the draft policy outlines measures to encourage innovation. These include the establishment of an AI Innovation Fund and the creation of regulatory sandboxes to allow AI systems to be tested in defined and supervised environments.
The policy also proposes integrating AI education into secondary, higher and technical curricula to help develop an AI-skilled workforce.
The document situates Bangladesh's AI governance within international norms, citing alignment with frameworks such as the OECD AI Principles, UNESCO's AI Ethics Recommendation, the Council of Europe's AI framework and ASEAN AI guidance.
The government said this alignment would facilitate international cooperation and responsible participation in global AI development.
The interim government has invited citizens, industry stakeholders, academics and civil society organisations to submit feedback on the draft policy through an online platform, stating that public input will be crucial to refining the framework and ensuring it reflects national priorities and values.
bdsmile@gmail.com