Eight connected pillars that decide whether artificial intelligence in Bangladesh is accountable, inclusive and fit for its context.
The AI governance ecosystem is everything that shapes how AI is built, bought, used and overseen in Bangladesh: policy and law, data, institutions, infrastructure, skills and the people affected. No single agency or document covers all of it.
The Bangladesh AI Governance Initiative, an independent research initiative of Atlas AI Institute, organises this ecosystem into eight pillars so that gaps are easier to see and fix.
Each card explains the pillar, why it matters in Bangladesh, and the questions the Initiative looks at.
Clear, proportionate rules and national strategies that can be put into practice, monitored and reviewed.
Why it matters: AI is being adopted faster than rules are written. Without workable policy, organisations either hold back or deploy with no guardrails.
Safeguards that protect dignity, rights and safety across the whole AI lifecycle.
Why it matters: AI can influence access to credit, jobs, services and information. Harms often fall first on people with the least power to challenge them.
Responsible stewardship of data: its quality, who can access it, and how personal information is protected.
Why it matters: AI runs on data. Personal data, public-sector data and Bangla-language data each need clear rules for collection, sharing and protection.
Oversight bodies with clear mandates, and redress routes that make AI answerable to the public.
Why it matters: Rules only work if someone is responsible for applying them and people know where to take a complaint.
Independent evidence and testing to understand what AI systems do and whether governance is working.
Why it matters: Policy built only on foreign evidence can miss local conditions. Bangladesh needs its own findings on how AI performs in Bangla and in local settings.
Connectivity, computing power and public digital systems that serve national priorities.
Why it matters: Whoever owns the infrastructure behind AI shapes who benefits and who carries the risk. Heavy reliance on foreign cloud and models raises questions of control, cost and resilience.
Building the people and public-sector expertise needed to use and oversee AI well.
Why it matters: Regulators, civil servants, judges, teachers and journalists need enough understanding to question AI systems, not just use them.
Context-specific guidance for sectors such as health, education, finance and agriculture, with no one left behind.
Why it matters: One rule cannot fit a hospital, a bank and a classroom. Inclusion also means Bangla speakers, rural communities, women and persons with disabilities.
The pillars are one system, not eight separate projects.
A rule with no regulator, mandate or redress route stays on paper.
Oversight depends on trained people, reliable data and usable infrastructure.
Research and evaluation show what works, so every other pillar can improve.
It is the set of laws, institutions, data, infrastructure, skills and stakeholders that together decide how artificial intelligence is developed, deployed and overseen in Bangladesh. The Bangladesh AI Governance Initiative organises it into eight connected pillars.
Six stakeholder groups: government and regulators, academia and research, industry and AI developers, civil society, citizens and communities, and international and regional partners.
Each pillar covers a different part of the system, and each depends on the others. Policy without institutions, skills or data cannot be put into practice, and none of it can be improved without evidence.
It is an independent research initiative of Atlas AI Institute.
Use the Contact page to reach the team for questions, collaboration and general enquiries.
Papers, frameworks and policy analysis, each with its authors, date and source.
