Bangladesh Has a National AI Policy — and It Is More Serious Than Most People Realise

On February 9, 2026, the ICT Division of Bangladesh published Draft Version 2.0 of the National AI Policy 2026–2030, the country's first comprehensive framework governing how artificial intelligence is developed, deployed, and regulated across its economy and public institutions. Public consultation closed on February 8, 2026, with the government reporting an "overwhelming response" from citizens, experts, and organisations across the country.

This is not a routine bureaucratic document. It is a signal about where Bangladesh positions itself as an emerging middle power in a global AI governance landscape that is being constructed in real time. And for students, job seekers, tech entrepreneurs, and anyone working in Bangladesh's digital economy, the policy's provisions on education, workforce development, industry regulation, and citizen rights have direct and practical consequences worth understanding.

The draft, published at the government's official AI policy portal, arrives at a moment when Bangladesh ranks 75th globally in the Oxford Insights Government AI Readiness Index — below India (which scores 0.49) but ahead of Pakistan (0.37) and Nepal (0.35), with a score of 0.38. Developed economies typically score 0.7 or above. The policy is, in part, an acknowledgment of that gap and a structured attempt to close it.

7 AI-Proof Careers That Will Survive 2026 and Beyond

The Context: Where Bangladesh Stands on AI Before This Policy

To understand what the policy is trying to do, it helps to understand what the landscape looked like without it. In November 2025, UNESCO, UNDP, and the ICT Division jointly released Bangladesh's AI Readiness Assessment Report — an evaluation conducted using UNESCO's Readiness Assessment Methodology, deployed in over 60 countries — and the findings were candid.

The report identified 15 priority actions and documented significant structural gaps: fragmented data systems, GPU scarcity, outdated university curricula, AI ethics instruction that is "nearly absent" from most programmes, and severe gender disparities in the AI workforce. Bangladesh's smartphone penetration rose by 72.8% in 2025 from 63% in 2023, and mobile-broadband subscriptions reached 57.8% of the population — which means the infrastructure for AI consumption exists, but the infrastructure for AI production remains substantially underdeveloped.

The policy also follows the failure of a previous attempt. The National Strategy for Artificial Intelligence, drafted in 2019–2020, included detailed roadmaps. Almost none materialised. Political disruptions account for part of that failure, but the absence of binding institutional commitments explains more. The 2026 policy drafters appear aware of this history — the document includes annual reporting requirements, a mandatory mid-term review in 2028, and a sunset clause requiring renewal by 2030. These accountability mechanisms are not standard in South Asian technology governance documents.

AI Skills That Pay 56% More: Complete Learning Guide 2026

What the Policy Actually Contains: The Key Provisions

The National AI Policy 2026–2030 is structured around several interconnected priorities. Understanding each requires separating what the document commits to from what it merely aspires to — a distinction that matters significantly for practical impact.

Risk-Based Regulatory Framework

The draft establishes a four-tier risk classification system for AI applications — prohibited practices, high-risk systems, limited-risk applications, and minimal-risk uses — that closely mirrors the architecture of the European Union AI Act. Explicitly prohibited under the draft: mass surveillance systems, social scoring of citizens by public authorities, and AI systems that manipulate human behaviour without consent. High-risk applications — AI used in employment, credit decisions, healthcare, education, and judicial contexts — would require mandatory Algorithmic Impact Assessments (AIAs) before deployment.

The commitment to ratify the Council of Europe's Framework Convention on AI would make Bangladesh potentially the first South Asian country to sign that instrument, signalling formal alignment with international accountability standards that go beyond voluntary principles. The policy also references the OECD AI Principles and aligns with UNESCO's Recommendation on the Ethics of AI — a framework adopted by all 194 UNESCO member states.

For students and citizens, this means: automated decisions that affect your education, job applications, loan approvals, or government service access would — if these provisions are implemented — come with explainability requirements, contestability mechanisms, and mandatory human review pathways. These are rights, not courtesies. Whether they will be enforceable is a separate and more complicated question addressed below.

Education and Workforce Development

The education provisions are where the policy speaks most directly to Bangladesh's young population. The draft commits to integrating AI, data science, and digital ethics into both school and university curricula — a structural reform that would affect every STEM graduate and eventually every secondary school student in the country.

Specifically, the policy directs investment in: establishing Centres of Excellence for AI expertise, ensuring inclusive access to AI education for women and underrepresented groups, integrating AI and data science across university curricula through the University Grants Commission, and supporting BTEB (Bangladesh Technical Education Board) to build AI vocational training programmes accessible to students outside the formal university system. The UNESCO Readiness Report, which fed into this policy, recommended creating "AI ambassador" networks at the local level to promote digital literacy — a bottom-up mechanism that complements the top-down curriculum reform.

The policy also explicitly prioritises Bangla-language AI development. This is a provision with real strategic significance: AI systems that work exclusively in English exclude the majority of Bangladesh's population from meaningful access. The draft states that AI models, datasets, and digital public services "shall place appropriate emphasis on Bangla and other nationally relevant languages" — a commitment to linguistic inclusion that, if implemented, would support both public access and domestic AI research capacity.

For students: if the curriculum integration provisions materialise on the committed 2026–2030 timeline, students entering university now or in the next two years can expect to see AI-related content becoming part of their standard coursework, regardless of whether they are in a technical or non-technical programme. The vocational training pathway through BTEB is equally significant for the majority of Bangladesh's young people who pursue trades and technical education rather than four-year university programmes.

Jobs and the Workforce

The policy's workforce development provisions reflect the tension between AI's productivity potential and its displacement risks — a tension that is particularly acute in Bangladesh given the country's large BPO sector, its 650,000+ freelancers, and the concentration of entry-level digital work that AI is most likely to affect.

The draft commits to nurturing "a skilled workforce that can utilise and build AI technologies" and to adopting data-driven policy-making in every government sector. But the UNESCO Readiness Report went further in its recommendations, specifically calling for a "gig and platform workers' rights Act" to ensure fair wages, social protections, and algorithmic accountability for Bangladesh's freelancers and digital workers. This recommendation, notably, does not appear as a concrete commitment in the draft policy — a gap that legal analysts have highlighted.

The ICT Division's parallel Digital Transformation Strategy commits to expanding the ICT workforce to seven to eight million professionals by 2030 and training 20,000 cybersecurity experts by 2027 and 50,000 by 2030. These are ambitious targets that require substantial and sustained institutional investment to achieve. The combination of AI policy and digital transformation strategy creates the clearest picture of what the government envisions: Bangladesh as a significant AI-capable workforce supplier to the global digital economy, not merely a consumer of AI products built elsewhere.

Tech Industry and Innovation

For Bangladesh's tech sector — the startups, software firms, fintech companies, and e-commerce platforms that currently represent the most active AI adoption in the country — the policy's industry provisions create both opportunity and obligation.

The draft commits to fostering a culture of AI research and innovation through public and private funding, establishing regulatory clarity (currently entirely absent), and positioning Bangladesh as a competitive player in regional and global AI governance. It also requires that AI systems developed or deployed in Bangladesh "account for cultural context, support local languages, and mitigate risks of linguistic or cultural marginalisation."

Private sector AI deployments classified as high-risk — which would include hiring algorithms, credit scoring systems, and customer service automation used in high-stakes contexts — would face the same AIA requirements as public sector systems. For companies currently deploying AI without formal assessments, this represents a compliance obligation that does not yet have an implementation date but will require preparation.

The policy also establishes an Independent Oversight Committee to monitor AI developments, with a multi-stakeholder composition including technical experts, ethicists, legal professionals, civil society, and representation from marginalised communities. This body, if functional, would be the primary accountability mechanism for the entire framework.

Top 10 Free AI Tools Every Student in Bangladesh Should Use in 2026

The Implementation Gap: The Honest Assessment

The draft's governance intentions are genuinely sophisticated by regional standards. Legal and policy analysts have been broadly positive about the framework's design while raising serious and specific concerns about implementation capacity.

The National Directorate of Government and Intelligence on AI (NDGIA) — the centralised government body that would coordinate the entire policy — does not yet exist as an operational institution. The Independent Oversight Committee requires an Act of Parliament that has not been drafted. The Ministry of Law is directed to "initiate the drafting of a comprehensive Artificial Intelligence Act by 2028" — which means binding law will not exist for at least two years after the policy's adoption, during which millions of Bangladeshis will continue interacting with AI systems without legal protections.

As one legal practitioner analysing the draft wrote for Medium, the gap between policy aspiration and legal enforceability "is where rights go to die." The policy lacks precise definitions of key terms, a clear roadmap linking policy provisions to existing laws, and concrete timelines for institutional establishment. The ICT Act 2006, the Personal Data Protection Ordinance 2025, the National Data Governance Ordinance 2025, and the Cyber Safety Ordinance 2025 collectively provide some interim statutory basis for action — but analysts argue regulators should be directed to act under these existing instruments immediately rather than waiting for a comprehensive AI Act that is years away.

Political continuity is the other variable. The policy was developed under the interim government led by Nobel laureate Muhammad Yunus, which oversaw the post-July Revolution transition period. Bangladesh's February 2026 elections introduced a new political context, and technology governance agendas in South Asia frequently reset during political transitions. The 2019–2020 strategy's failure is a precedent that anyone assessing the 2026 policy's prospects should take seriously.

How Bangladeshi Freelancers Are Using AI to Earn More on Upwork and Fiverr in 2026

What This Means in Practice: Three Audiences

For students and young people: The most immediately relevant provision is the curriculum integration commitment. Students in computer science, engineering, and business programmes should expect AI-related content to be formalised in their coursework over the 2026–2028 period. The BTEB vocational pathway matters equally for technical education students. Beyond formal curricula, the policy signals that the government intends AI literacy to be a national priority — which means the tools, courses, and resources for self-directed AI learning will likely receive more institutional support than they have previously.

The rights provisions — explainability, contestability, human review — matter for every student who will eventually navigate automated hiring systems, university admission algorithms, or government service platforms. Knowing these rights exist, and knowing how to invoke them, is part of AI literacy that the policy is designed to make universal.

For working professionals and job seekers: The workforce scale-up targets — seven to eight million ICT professionals by 2030, 50,000 cybersecurity experts — signal where government investment and training infrastructure will be directed. The skills that the policy positions as strategic priorities — AI, data science, cybersecurity, digital ethics — align closely with the global wage premium data covered in our analysis of AI skills and earning power in 2026. This policy does not create opportunities by itself, but it does signal where institutional infrastructure and public investment will concentrate.

For tech entrepreneurs and companies: The clearest near-term implication is regulatory direction. Companies deploying AI in high-risk categories — employment screening, credit decisions, healthcare, education — should begin developing internal AIA processes now, before binding requirements take effect. Companies building Bangla-language AI capabilities, developing AI for agriculture or rural public services, or focusing on areas the policy identifies as priorities are likely to be better positioned for future public procurement and partnership opportunities.

Bangladesh in South Asia's AI Race

Bangladesh's AI policy arrives as every major South Asian economy is simultaneously staking out AI governance positions. India hosted the AI Impact Summit 2026 in New Delhi in February, positioning itself as the voice of developing economies on global AI governance. Sri Lanka (AI readiness score 0.44) and India (0.49) both score above Bangladesh on the Oxford Insights index. Pakistan (0.37) and Nepal (0.35) score below.

The commitment to ratify the Council of Europe's AI Convention, if followed through, would be a meaningful regional differentiator. It would make Bangladesh the first South Asian country to formally align with international AI accountability standards that include binding obligations, not just voluntary principles. Whether this happens depends on political will that survives institutional transition — which, given South Asia's track record with technology governance commitments, is genuinely uncertain.

What is less uncertain is that the process Bangladesh has followed — the UNESCO RAM assessment, the multi-stakeholder consultation, the engagement with international frameworks — represents more methodical governance preparation than most comparable economies have undertaken. The framework exists. The question is whether institutions will be built to implement it.

For more on how AI skill development connects to career opportunities in Bangladesh's evolving digital economy, see our guide to which careers offer the strongest structural resilience as AI reshapes the labour market. As the final version of the policy is adopted and implementation begins, WinTK will continue tracking developments and their practical implications for Bangladesh's workforce and tech sector.