Published: 12:49 AM, 24 November 2025
Millions of Jobs at Risk: How AI Could Widen Inequality in Bangladesh
Albert Jackson Biswas
What happens when a university degree becomes a liability instead of an asset? In Bangladesh today, graduates face 13.54 percent unemployment, while people with no formal education experience just 1.25 percent joblessness. This isn't a statistical anomaly. It's a warning sign of a technological transformation that could reverse three decades of inclusive development within just three years.
A 32-year-old professional working remotely from Dhaka for a UK-based company discovered this firsthand. The arrangement once seemed like the perfect blend of global opportunity and staying close to home. But over the past year, she's watched her company's Bangladesh operations shrink dramatically. The Dhaka branch that once employed 133 analysts now has 70. The reason isn't offshoring or cost-cutting in the traditional sense. It's AI and Business Intelligence systems that can perform in seconds what took her colleagues hours. "They're not firing people," she explains quietly. "They're just not replacing anyone who leaves. The algorithms are doing that instead."
Her story isn't unique. It's becoming Bangladesh's new reality.
The unemployment rate climbed to 4.63 percent in the final quarter of 2024, up from 3.95 percent a year earlier, a seemingly modest increase that masks a deeper structural transformation reshaping the country's economic future. Behind these numbers lies a technological shift that economists warn could reverse decades of inclusive development gains.
The country's Gini coefficient, which measures income inequality on a scale from 0 (perfect equality) to 1 (maximum inequality), reached 0.50 in 2022 according to the White Paper on the State of Bangladesh Economy. This places Bangladesh at a critical threshold. Conservative projections suggest the coefficient could rise to 0.53-0.58 by 2028, pushing the country toward inequality levels typically associated with Latin America's most divided societies, where the wealth gap resembles a chasm rather than a curve.
This U-shaped unemployment pattern, where joblessness is low at both extremes but high in the middle, reveals precisely how artificial intelligence affects different types of work. AI excels at routine cognitive tasks: processing loan applications, analyzing risk, managing compliance, and generating reports. These are precisely the jobs that required university degrees and promised middle-class stability, the positions that parents sacrificed to help their children attain. Meanwhile, AI struggles with non-routine manual work and advanced problem-solving. A construction laborer's job remains relatively secure. A senior executive's strategic role stays safe. But the junior analyst, the compliance officer, the credit assessor? They're becoming obsolete.
Political instability in 2024 compounded the employment crisis through multiple channels. Women's unemployment surged to 7.16 percent from 6.15 percent year-over-year, with security challenges in urban areas limiting workforce participation. For a country where women's economic engagement has been the engine of poverty reduction for two decades, this reversal threatens not just individual families but Bangladesh's entire development model.
Economic modeling based on observed automation rates suggests troubling trajectories for the immediate future. Under conservative assumptions, with routine task automation proceeding at 3-5 percent annually and limited educational adaptation, unemployment could reach 6.0 percent by 2028 in the best case. In pessimistic scenarios, where political instability compounds technological displacement, it could hit 9.0 percent.
Youth unemployment, currently at 11.46 percent, faces particularly acute pressure. Projections suggest rates could climb to 15-18 percent by 2028, with graduate unemployment potentially reaching 17-21 percent. These describe conditions affecting young people entering the workforce right now, students currently in their final university years, planning careers that may not exist by the time they graduate.
The ready-made garment sector offers empirical evidence of what's already happening. A Bangladesh Labor Foundation study documented a 30.58 percent employment reduction following technological upgrades in the sector employing 4 million workers, 80 percent of them women. With garments generating 84 percent of export earnings, even modest displacement creates ripple effects throughout the economy. When garment workers lose income, the tea stall owner loses customers, the tuition teacher loses students, and the local shop sees sales decline.
The banking sector, employing 203,696 people across 61 commercial banks as of December 2023, according to Bangladesh Bank, exemplifies white-collar vulnerability. International institutions have deployed AI for Expected Credit Loss calculations, Risk-Based Supervision, and IFRS 9 compliance, functions that once employed thousands of Bangladeshi professionals in routine analytical roles. Walk into any bank from Motijheel to Agrabad, and beneath the professional courtesy, you'll find the same quiet anxiety.
The skills mismatch creates a trap: banks need fewer junior analysts but more data scientist roles requiring technical training that fewer than 5 percent of current banking professionals possess. It's not enough to be willing to learn. The transition requires years of technical education that most can't afford and institutional capacity that doesn't yet exist at scale.
Think of inequality as a curve. Economists use the Lorenz curve to visualize it, plotting cumulative income share against cumulative population share. Perfect equality appears as a diagonal line. Actual distribution bows below it. The further the bow, the greater the inequality. Bangladesh's curve, already significantly bowed at Gini 0.50, could deepen substantially by 2028, bending toward the extreme inequality that characterizes Latin America's most divided societies.
This matters because inequality affects not just distribution but growth sustainability. Bangladesh's Human Development Index reached 0.685 in 2023, up from 0.397 in 1990, a 72.5 percent increase, placing the country among Asia's best performers. But rising inequality undermines these gains. The Inequality-adjusted HDI already reduces Bangladesh's headline figure from 0.670 to 0.470 due to existing disparities.
Political instability following August 2024's transition compounded technological challenges through multiple channels. Foreign investment declined, slowing job creation in sectors that might absorb displaced workers. Domestic business confidence erosion delayed expansion plans, reducing labor demand precisely when automation increases supply. Industries faced challenging conditions with both labor supply and demand decreasing significantly, creating unemployment dynamics more severe than either factor alone would produce. The timing proves particularly unfortunate. Bangladesh requires activist labor market policies, massive retraining investments, social protection expansion, and education system overhaul precisely when institutional capacity faces transition constraints. Government education spending remains at just 1.53 percent of GDP in FY 2025-26, well below UNESCO's 4-6 percent recommendation.
Several policy interventions could alter these trajectories, but the window for action narrows with each passing month. Education system transformation represents the most crucial lever. Massive investment in technical education, with curricula aligned to AI-complementary skills, could convert demographic challenge into an advantage. Singapore's SkillsFuture program, providing lifelong learning credits to every citizen, offers an adaptable model. Vietnam's electronics sector transition demonstrates how strategic industrial policy can create demand for educated workers while maintaining export competitiveness.
Social protection expansion could cushion displacement while maintaining human capital. Unemployment insurance and retraining stipends prevent forced shifts into informal low-productivity work that reduce lifetime earnings. Progressive taxation, potentially including levies on automation's productivity gains, could fund redistribution without excessive distortion if designed carefully. The interim government's November 2024 establishment of a labor reform commission signals institutional recognition, though comprehensive policy responses remain emergent. International development partners could accelerate transitions through technical assistance and concessional financing for workforce development.
Bangladesh confronts a choice that will define its development trajectory for decades. The country can leverage AI for productivity gains that raise living standards broadly, or allow market forces alone to concentrate gains narrowly while displacing millions. Technology itself amplifies human choices about distribution, magnifies institutional capacity or its absence, and accelerates both progress and inequality.
The 203,696 banking sector workers, 4 million garment employees, and millions more in vulnerable sectors represent not just employment statistics but families, communities, and Bangladesh's inclusive growth model built painstakingly over three decades. Whether the country's curve continues bowing toward severe inequality or bends back toward shared prosperity depends on decisions made now. For professionals like her and millions of others, the question isn't whether AI will transform their work. It's whether that transformation will expand opportunity or eliminate it entirely.
Bangladesh has overcome greater challenges through determined collective action. The devastating floods. The global financial crisis. Each time, collective determination and strategic action produced resilience. This challenge requires similar determination, but the window for effective action closes rapidly. The forecasts showing Gini coefficients reaching 0.53-0.58 by 2028 aren't predictions cast in stone. There are warnings about what happens without intervention. The economic theory is clear. The international evidence supports it. The policy tools exist and could be adapted to Bangladesh's context. What remains is political will deployed at sufficient scale and speed.
By 2028, today's university freshmen will be entering a job market that will either offer opportunities or deny them systematically. The young professional working remotely will either see expanding possibilities or watch her entire sector collapse. These aren't abstract scenarios. They are the lived reality that will unfold based on choices made in the next eighteen months.
The window remains open, but it's closing fast. Not gradually over decades, but rapidly over quarters. The question isn't whether Bangladesh has the resources or knowledge to respond effectively. The question is whether it has the political will to act before the trajectory becomes irreversible. For millions of Bangladeshis, that answer will determine everything.
Albert Jackson Biswas is a certified
data scientist & AI engineer.
