Bangladesh National AI Policy 2026-2030: The Government Roadmap, What It Promises, and What Could Go Wrong

A Country That Has Written the Strategy Twice

Bangladesh has now produced two major AI policy documents. The first, the National Strategy for Artificial Intelligence, was released in March 2020 by the Information and Communication Technology Division under the government of Sheikh Hasina. Its slogan was "AI for Innovative Bangladesh." It identified six strategic pillars, seven priority sectors, and an ambitious five-year roadmap. It promised to train 50,000 AI professionals, establish 1,000 AI startups, and integrate artificial intelligence into governance, agriculture, healthcare, education, manufacturing, transportation, and finance. It was, on paper, a serious document for its time — one of the earliest comprehensive AI strategies produced by a developing country in South Asia.

Nearly six years later, the strategy's own critics describe it as largely aspirational. No concrete implementation plans followed. No institutional mechanisms were established to own the implementation. Budget allocations did not materialize at the scale the document envisioned. Political disruptions played a role — but the absence of binding commitments explains more than the politics does. The 2020 strategy did what many national AI strategies do: it framed the ambition, then left the building to happen on its own.

Now Bangladesh has a second attempt. The National AI Policy 2026-2030, published by the ICT Division in draft form and currently accepting stakeholder feedback, is a more sophisticated document than its predecessor — more specific on governance, more explicit on rights and risks, and produced in a fundamentally different political context. The interim government of Nobel laureate Professor Muhammad Yunus, which came to power following the July 2024 uprising that removed Sheikh Hasina, has made integrating AI into the national development agenda a stated priority. The policy aligns this ambition with Vision 2041 and the Sustainable Development Goals. Whether this draft becomes the implementation plan the 2020 strategy never had is the question Bangladesh's technology sector is now asking.

Top 10 AI Companies in Bangladesh 2026

What the 2020 Strategy Actually Said

The original national AI strategy was a landmark document for Bangladesh. Commissioned by the a2i (Aspire to Innovate) programme and executed by HyperTAG Solutions Ltd, authored by national AI strategist Md Shofiul Alam, it laid out a structured analysis of Bangladesh's AI readiness and a roadmap for closing the gap between where the country was and where it needed to be.

The six strategic pillars the document identified have remained the reference framework for Bangladesh's AI conversations ever since: research and development; skilling and reskilling the AI workforce; data and digital infrastructure; ethics, data privacy, security, and regulations; funding and accelerating AI startups; and industrialization and commercialization of AI technologies. Seven priority sectors were identified — public service delivery, manufacturing, agriculture, smart mobility and transportation, skill development and education, finance and trade, and healthcare.

The document was grounded in the Digital Bangladesh context: ICT export earnings had grown from $26 million in 2008 to nearly $1 billion, sixteen Hi-Tech Parks had been built, a Tier-IV Data Center was operational, and the Bangabandhu-1 satellite had placed Bangladesh in the space sector. The argument was that AI would be the next accelerator — that the digitization Bangladesh had already achieved created a foundation for AI adoption that the country should now build on strategically.

The economic case the strategy made was compelling: Accenture research cited in the document argued that AI has the power to double annual economic growth rates by 2035 in developed economies, and that labor productivity can be increased by up to forty percent through AI-enabled efficiency. For Bangladesh — a labor-intensive economy facing middle-income transition — the appeal of AI as a productivity driver was self-evident. What the strategy could not fully answer was the institutional question: which body would own the implementation, with what budget, accountable to whom?

The Gap Between Strategy and Reality

By the time Bangladesh's political landscape shifted dramatically in July 2024 — when the student-led uprising removed Sheikh Hasina's government — the 2020 AI strategy had produced limited concrete results. The target of 50,000 trained AI professionals had not been reached through dedicated AI-specific programming. The 1,000 AI startups target was aspirational against a backdrop of a startup ecosystem that, while growing, numbered around seventy-three AI companies in Dhaka according to Tracxn data — significant growth, but not at the scale the strategy envisioned. The national AI research center the document called for had not been established. The AI sandboxes had not materialized in any formal, operational sense.

Bangladesh is not alone in this experience. A widely cited analysis of national AI strategies globally found that implementation rates — the proportion of strategy commitments that translate into concrete programs, legislation, or institutional changes — are consistently lower than the strategies' rhetoric suggests. India's NITI Aayog, Rwanda's regulatory authority-led approach, and Indonesia's regulatory sandbox model are frequently cited as examples of countries that moved from strategy documents to institutional ownership faster and more effectively. Bangladesh ranked 75th globally in the Oxford Insights Government AI Readiness Index, a position that reflects the gap between policy ambition and operational infrastructure.

The honest assessment, published in Bangladesh's own academic and policy press, is that the 2020 strategy was a crucial first step that could not be the last. It built a shared vocabulary and a shared framework. It identified the right sectors and the right challenges. What it could not do was create the institutional backbone that implementation requires — and without that backbone, no strategy document, however well-written, produces transformation.

Digital Copyright and Content Creator Rights: DMCA Lessons for Bangladesh Streamers

The 2026-2030 Draft: What Is Different This Time

The National AI Policy 2026-2030, currently in its second draft as of February 2026, reflects both the lessons of the previous strategy's limitations and the changed political priorities of the Yunus government. Several features distinguish it from the 2020 document in ways that matter for implementation.

First, the governance architecture is more explicit. The policy proposes a National Digital Governance and Intelligence Authority (NDGIA) as the central coordinating body for AI implementation across government. The ICT Division would retain overall policy direction, with line ministries required to align their AI initiatives with ICT Division guidance to avoid duplication of systems and costs. This centralization is deliberate: one of the key failure modes of the 2020 strategy was diffuse ownership, where every ministry could claim AI relevance but none was accountable for delivering specific outcomes.

Second, the rights and risk framework is substantially more developed. The 2026-2030 draft includes a risk-based classification system with four tiers, from prohibited practices through high-risk applications to limited-risk and minimal-risk use. Explicit prohibitions on mass surveillance and social scoring are included — a notable commitment given Bangladesh's historical experience with surveillance infrastructure. Mandatory algorithmic impact assessments for high-risk systems are required. The policy commits Bangladesh to ratifying the Council of Europe's Framework Convention on AI, signaling an alignment with international governance norms that the 2020 document did not make explicit.

Third, the digital sovereignty dimension reflects the new government's priorities in a way the 2020 document did not address. A cornerstone commitment in the 2026-2030 draft is the development of a Bangla-based large language model — a national AI system analogous to ChatGPT or Gemini but built on Bangla language and cultural context, intended to preserve linguistic heritage, make AI contextually relevant, and protect intellectual property from foreign exploitation. The government will restrict the use of unsecured, API-dependent foreign AI systems and prioritize domestic hosting of secure, Bangladesh-specific language models. Citizen redress mechanisms will allow individuals to challenge AI-driven decisions that affect their rights or access to services.

Fourth, the accountability mechanisms are more specific. Annual reporting requirements are included. A mandatory mid-term review in 2028 is required. A sunset clause mandates renewal by 2030, when the policy is intended to be replaced by a permanent Artificial Intelligence Act. These are accountability cycles that many national AI strategies lack entirely — they create pressure for implementation that self-referential strategy documents cannot generate.

Priority Sectors: Where AI Is Supposed to Go First

The 2026-2030 policy prioritizes the same broad sectors as the 2020 strategy — agriculture, healthcare, public services, finance, education — but with more operational specificity about what AI is actually supposed to do in each.

Agriculture gets the most concrete treatment. AI applications in this sector will support precision irrigation, pest detection, and localized weather forecasting for Bangladesh's sixteen million farming households. Agriculture employs forty percent of Bangladesh's workforce but contributes only eleven percent to GDP — a productivity gap that AI-driven precision farming has shown, in pilot programs, can be addressed with forty percent reductions in crop losses. Intelligent Machines, one of Bangladesh's established AI startups, is already working in this space. The policy creates a framework for scaling what currently exists in pilot form into national deployment.

Healthcare focuses on public health management and crisis prediction. The policy is careful — life-altering clinical decisions remain under certified medical professionals — but AI-assisted diagnostics, telemedicine support, and predictive epidemiology are explicitly within scope. Bangladesh has one doctor for every 900 people, a ratio that makes the case for AI-assisted healthcare delivery more urgent here than in most developed economies.

Public service delivery is where the immediate and visible wins are expected. The strategy's phase one targets — 2025 to 2026 — include the launch of the Bangladesh National Digital Architecture (BNDA), a National Data Exchange (NDX), and the digitization of more than 800 government services. The goal within six years is to reach the top fifteen in the UN E-Government Development Index. Bangladesh jumped from 119th to 100th in global e-government rankings in four years — already the top performer among Least Developed Countries — which gives the target at least some credibility.

Finance and fintech follow from the infrastructure bKash built. With seventy million mobile financial services users, Bangladesh has more behavioral data on small-scale economic transactions than most comparable economies. AI-driven credit scoring for the unbanked, micro-lending analytics, fraud detection — the 2026-2030 policy creates a framework for deploying AI against a financial inclusion agenda that already has the data infrastructure to support it.

The Smart Bangladesh 2041 Context

The AI policy sits within a larger strategic framework — Bangladesh's Perspective Plan 2021-2041, which frames Vision 2041 around eliminating extreme poverty and reaching upper-middle income status by 2031. The AI policy is explicitly positioned as one of the accelerators for this transition. The National Digital Transformation Strategy, unveiled in partnership with UNDP in February 2025, targets a $5 billion ICT export sector by 2030, 7 to 8 million ICT professionals, a fifty percent increase in startup funding, and for Bangladesh to become the AI and Fourth Industrial Revolution hub in South Asia.

These numbers are ambitious against current baselines. Bangladesh's ICT exports are currently around $1 billion. The digital literacy rate is eight percent. Mobile internet speed averages 9.2 megabits per second against a global average of 64.2. The talent pipeline — 650,000 freelancers, BUET and BRAC University AI programs, the Learning and Earning Development Project's training of 100,000 freelancers — is real but not yet at the scale the 2030 targets require. The gap between strategy and outcome, which defined the 2020-2026 period, remains the central challenge.

The Institutional Question: NDGIA and Political Continuity

The most important uncertainty around the 2026-2030 policy is institutional rather than technical. The NDGIA — the central coordinating body the policy proposes — does not yet exist as an operational institution. The policy references it throughout without an establishment timeline or capacity benchmarks. The Independent Oversight Committee the policy calls for requires an Act of Parliament that has not been drafted.

Bangladesh has been in this position before. The 2020 strategy also called for institutional creation — an AI Task Force, coordinating mechanisms, multi-stakeholder governance — and those institutions either did not materialize or did not function with the authority the strategy required. The difference, policy analysts argue, is that the 2026-2030 draft includes accountability mechanisms the 2020 document lacked: the mid-term review, the sunset clause, the annual reporting requirements. But mechanisms are only as strong as the political will to use them.

The political transition question is real. Bangladesh held its first election since the July Uprising in February 2026. Political transitions in South Asia frequently reset technology governance agendas — new governments inherit strategy documents but not always the institutional momentum behind them. Whether the incoming elected government continues the AI policy work begun under the Yunus interim administration, and with what institutional commitment, is a question the policy document itself cannot answer.

What It Means for Bangladesh's Tech Workers and Students

For the people most directly affected by this policy — the students at BUET studying machine learning, the software engineers in Gulshan considering an AI specialization, the freelancers building AI skills on platforms like Coursera and Udemy, the startup founders trying to build AI products in a market with limited domestic venture capital — the practical implications of the 2026-2030 policy are several.

The government's commitment to a national Bangla LLM creates a potential anchor project for AI talent in Bangladesh — a large, technically serious program that requires the kind of AI engineering capacity the country is still building. The skilling commitments, if implemented, create pathways from general software development into AI specialization. The regulatory framework — risk-based, with algorithmic impact assessments for high-risk systems — creates compliance requirements that AI companies will need legal, technical, and operational expertise to navigate.

The $5 billion ICT export target by 2030 is not achievable without AI. Bangladesh's current export mix — predominantly software outsourcing and business process management — does not reach that figure without moving up the value chain. AI services, AI-enabled products, and the data annotation and labeling work that feeds global AI systems are where that additional revenue has to come from. The policy creates the framework. Whether the institutions, the infrastructure, and the political continuity exist to make the framework functional is what the next five years will determine.

win-tk.org is a WinTK publication.