Bangladesh is preparing for an AI economy, but its existing technology base offers a warning that must be heeded. The Economic Census 2024 found that only 2.44 percent of manufacturing establishments in the country use computers or information technology in production. The FY2026-27 budget, meanwhile, projects a vision of specialised laboratories for AI, machine learning, and big data, alongside wider use of AI in public services, education, health, and agriculture. Before forging ahead towards that technicolour future, some primary questions should be answered: who will be able to turn AI into higher productivity, and how will those benefits accrue across our society?
Bangladesh’s first major economic transformation succeeded because a large labour force was connected to global production. Ready-made garments became the clearest expression of that model. The next challenge is different. Between 2016 and 2022, 1.4 crore young Bangladeshis became working age adults while the economy added only 87 lakh jobs. Nearly 70 percent of those who found work entered low-productivity agriculture. Bangladesh needs more jobs, but it also needs to raise workers’ and firms’ productivity.
AI changes both sides of this problem; it leads to new capabilities while some existing sources of advantage, such as low-cost labour, lose value. Labour-saving technologies can make low labour costs less decisive in investment decisions. The next sunset industry may not travel to the Global South the way earlier labour-intensive industries did; some of that work may disappear as firms automate rather than relocate.
AI is also making some forms of useful knowledge dramatically cheaper. Stanford’s AI Index estimates that the cost of using a model performing at roughly GPT-3.5 level fell from about $20 per million tokens in November 2022 to $0.07 by October 2024. That matters because useful knowledge has never been evenly distributed across firms. Large companies can employ people to research markets, analyse accounts, manage compliance, develop software, translate documents and advise management. Smaller firms usually cannot maintain all of these capabilities internally or buy them regularly from specialists. However, thanks to increasingly capable AI tools, knowledge that was once expensive to organise may become available much further down the economy, benefiting most Bangladeshi firms as well.
But cheaper knowledge does not automatically become broader capability. Productivity in Bangladesh is highly uneven across sectors and firms. World Bank research shows that Bangladesh’s most productive firms are about 11 times as productive as other formal firms and account for around 70 percent of exports, yet only about 15 percent of formal employment. Firms already ahead also tend to have better skills, capital and systems with which to use new technology. Besides, smaller suppliers will face buyers expecting faster turnaround, better traceability or more sophisticated reporting as competitors adopt AI, even though they may lack the capacity to respond similarly. Meanwhile, CEOs will increasingly talk about replacing “low-quality human capital” with AI.
If AI mainly makes firms that are already productive even more productive, Bangladesh can become more technologically advanced without solving its deeper jobs and productivity problem. This would again feed back into the loop of stagnating growth. The larger AI opportunity should be for cheaper knowledge and capability to reach the smaller firms.
Inclusion should mean more than AI access. Smaller firms do not need to become AI companies; they need support to apply existing tools to real business problems and share what works. Universities, industry associations and business-support organisations can help firms test uses in market research, compliance, production planning and accounts. Meanwhile, the government can help improve the conditions for productive use—practical skills, Bangla resources, usable public information and research capacity.
The FY2026-27 budget itself calls for “democratising economic opportunities” and defines inclusive development in terms of meaningful participation and broadly shared gains. Much depends on how this is implemented. For each investment, the test should be simple: who becomes more capable because of it?
Bangladesh’s first transformation broadened participation in a growing productive economy. AI presents a different technology and a different moment, but the development test is similar. If cheaper useful knowledge mainly makes the strongest firms and institutions stronger, Bangladesh may adopt AI without broadening economic opportunity. However, if workers and firms that start with fewer advantages can use AI to become more capable, the productive base will widen, and a foundation for better jobs and the next stage of growth will be created.
How Bangladesh uses AI can be a nebulous question. The more immediate, basic question is whether AI will help broaden the part of the economy that can produce, compete and create better jobs, or whether it will worsen the inequality problem.
Saba El Kabir is founder of Cultivera, and institutional sustainability adviser at PPRC. He can be reached at [email protected].
Views expressed in this article are the author's own.
Follow The Daily Star Opinion on Facebook for the latest opinions, commentaries, and analyses by experts and professionals. To contribute your article or letter to The Daily Star Opinion, see our guidelines for submission.
