AI challenges Bangladesh's promise of inclusion

For more than a decade, 'Digital Bangladesh' was not merely a political slogan; it functioned as a practical framework for governance, service delivery, and socio-economic transformation. Internet connectivity expanded rapidly, government services became digitised, and millions of citizens-many of them poor and rural-were brought into the formal economy for the first time through mobile financial services. Technology, despite its limitations, carried a relatively inclusive promise. As Bangladesh now moves toward a new ambition often described as 'AI Bangladesh,' the stakes have become far higher. The central question is no longer about access alone, but about power, exclusion, and discrimination. Who will benefit from this transformation, and who will quietly fall behind?

Relatively low barriers to entry characterised the early phase of Digital Bangladesh. A basic smartphone and affordable mobile internet were often enough to participate in the digital economy. According to the Bangladesh Telecommunication Regulatory Commission, by October 2025, the country had around 132.5 million mobile internet subscribers. Mobile Financial Services accounts exceeded 125 million, with average daily transactions of roughly Tk 4,500 crore. These platforms opened unprecedented avenues of financial inclusion, particularly for the rural poor and low-income households. In the global freelancing market, Bangladesh emerged as the second largest supplier of online labour, with around 650,000 active workers. While these benefits were unevenly distributed, there were few institutional or technological barriers that actively excluded large segments of the population.

Artificial intelligence fundamentally alters this inclusive equation. AI is not simply an extension of digital connectivity. It relies on high-performance computing infrastructure, large volumes of structured data, advanced technical skills, and strong institutional capacity. The pathway to participation in 'AI Bangladesh' is therefore far narrower and far more expensive than that of Digital Bangladesh. This shift risks transforming technology from an enabler of inclusion into a driver of exclusion.

Data from DataReportal's 2025 annual report illustrates this concern clearly. While internet usage in Bangladesh has reached about 47.4 per cent of the population, only 16 per cent of households have access to high-speed broadband internet. In effect, more than half of the country's people remain on the margins of the digital world, relying on low-quality or unstable connections. If AI-enabled services in education, healthcare, finance, and public administration are built on top of this existing digital divide, they will inevitably favour a small, affluent, and urban segment of society. Those who are already ahead technologically will accelerate further, while those left behind may face a deepening and more permanent form of inequality.

The labour market is where the disruptive potential of AI becomes most visible. Around 84.9 per cent of Bangladesh's total employment is concentrated in the informal sector, where work is often repetitive, low-skilled, and vulnerable to automation. The International Labour Organisation estimates that roughly 54 per cent of occupations in Bangladesh will undergo some degree of transformation over the next decade due to AI and automation. This does not necessarily imply mass unemployment, but it does signal profound changes in job content, skill requirements, and employment security.

Globally, advanced economies are already boosting productivity through robotics and AI-driven software. Bangladesh is following this trajectory, albeit unevenly. AI applications are increasingly visible in manufacturing, transport management, retail operations, logistics, and even routine office work. While this may enhance efficiency and competitiveness, it also threatens millions of low-skilled workers. Occupations such as driving, data entry, and courier services are particularly exposed. Without viable pathways for reskilling and alternative employment, the country risks a surge in structural unemployment. Research by the Bangladesh Institute of Development Studies shows that around 33 per cent of educated unemployed individuals already lack the appropriate digital skills to secure jobs-a gap that is likely to widen in the AI era.

The readymade garment sector, the backbone of Bangladesh's export economy, provides a stark illustration of these dynamics. Large and modern factories have begun adopting automatic cutting machines, AI-based quality control systems, and smart production management tools. According to the Centre for Policy Dialogue, over the past decade, increased mechanisation in the garment sector has been associated with a decline of about 22 per cent in employment intensity for every additional unit of investment. While productivity and compliance standards improve, employment opportunities-especially for low-skilled workers-shrink.

This trend has serious gender implications. Nearly 60 per cent of the garment sector's approximately four million workers are women. Without a coherent national strategy for technical upskilling and workforce transition, AI risks creating a new hierarchy within the sector itself. Highly skilled technicians and engineers may see wage increases of 30 to 40 per cent, while ordinary sewing operators face job losses and wage stagnation. This is not merely an economic issue; it carries profound social consequences for female labour force participation, household welfare, and urban poverty.

The rural-urban divide further complicates the AI transition. While mobile internet coverage now reaches almost all parts of the country, broadband speed and stability remain weak in rural areas. According to the Bangladesh Bureau of Statistics ICT Usage Survey 2024, only 3.8 per cent of rural residents use a computer or laptop, compared to 19.5 per cent in urban areas. AI-powered solutions in education, healthcare, and agriculture may appear promising in policy documents, but in practice, they risk becoming urban-centric. A marginal farmer without access to AI-based weather forecasting, soil analysis, or market intelligence tools cannot compete with larger, technologically enabled agribusinesses. Over time, this digital divide could translate into widening income and opportunity gaps between rural and urban populations.

Education stands at the core of this challenge. AI-driven economies demand skills such as problem-solving, analytical reasoning, adaptability, and ethical judgment, rather than rote memorisation. The World Economic Forum estimates that AI-related technical competencies will account for around 44 per cent of required job skills by 2030. Yet Bangladesh's education system remains poorly aligned with these demands. While students in elite English-medium schools and expensive private universities increasingly gain exposure to coding, data science, and AI tools, around 75 per cent of students in government schools and madrasas lack access to adequate laboratories, trained teachers, or even reliable internet connectivity.

This educational divide risks producing a small 'AI elite' that dominates high-value jobs, corporate leadership, and policy-making, while the majority of young people are locked out of future opportunities. Talent alone will not be enough; without access and institutional support, capable students from ordinary families may be excluded from the AI-driven economy.

Gender inequality adds yet another layer of risk. In Bangladesh, women's internet usage and smartphone ownership rates are 29 per cent and 21 per cent lower than those of men, respectively. A GSMA report identifies Bangladesh as having one of the largest gender gaps in mobile internet usage in South Asia. At present, there is no comprehensive national initiative aimed specifically at increasing women's participation in advanced technology and AI research. Without targeted policies, AI may reproduce and even amplify existing gender disparities rather than serve as a tool of empowerment.

Older citizens are also being overlooked in this technological transformation. Senior citizens account for around 8 per cent of Bangladesh's population, yet many struggle with digital interfaces in banking, healthcare, and government services. As automation and algorithmic decision-making expand, there is a real risk that technological complexity could translate into the denial of basic rights. Efficiency gains must not come at the expense of accessibility, accountability, and human dignity.

AI's dependence on data introduces further concerns. Data collection in Bangladesh remains fragmented, poorly regulated, and heavily urban-biased. Marginalised groups are often underrepresented or absent from official datasets. When AI systems are used for decisions such as loan approvals, job screening, or welfare targeting, these data gaps can result in algorithmic bias-systematic discrimination embedded in mathematical models. Moreover, heavy reliance on foreign AI platforms raises questions of data sovereignty and national security. Without developing local AI models that reflect the Bangla language, culture, and social realities, Bangladesh risks long-term technological dependency.

Artificial intelligence undoubtedly offers enormous potential for Bangladesh, from disaster prediction and healthcare diagnostics to efficient governance and productivity growth. However, realising this potential requires a coordinated and inclusive strategy. Education reform must prioritise hands-on learning, coding, and AI ethics over rote memorisation. Infrastructure investment must focus on narrowing the urban-rural broadband gap in both speed and affordability. Large-scale reskilling and transition funds are essential for workers at risk of displacement. Strong data protection laws and ethical guidelines must safeguard citizens' rights. Above all, gender inclusion and support for marginalised groups must be embedded at the heart of AI policy.

The transition from Digital Bangladesh to AI Bangladesh is not simply a technological upgrade; it is a deeply social and political choice. If AI-driven growth enriches only a small, privileged class while rendering millions unemployed or irrelevant, it will lack any moral foundation. Development has meaning only when it improves the lives of those at the lowest rungs of society. AI must not become a glamorous slogan detached from social reality, but a tool for empowering every citizen. The greatest challenge facing Bangladesh is to ensure that no one becomes outcast or invisible in this wave of technological change. Only by placing human dignity and social justice at the centre of AI adoption can Bangladesh truly build a resilient and inclusive future.

Dr Matiur Rahman is a researcher and development professional.

matiurrahman588@gmail.com