Making AI work for Bangladesh - The Financial Express

TEXT START: In its latest World Development Report, the World Bank asserts that Bangladesh could compress a century of development into a single decade by harnessing artificial intelligence.

The Dissection

The article performs developmental reassurance. It recasts AI as a cheap productivity accelerator that can help Bangladesh catch up, while treating labour displacement as a manageable side effect requiring retraining.

Its strongest observation is that AI initially targets tasks rather than entire occupations, and that entry-level work is especially exposed. But the article treats this as a warning about workforce pipelines, not as evidence that the wage-to-consumption system itself is being hollowed out.

The garment sector is presented as temporarily protected because human labour remains cheaper than machinery. That is not a durable advantage. It is a lag defense based on current wages, infrastructure, and capital costs. Once AI-assisted inspection, scheduling, quality control, and production systems become cheaper and standardized, Bangladesh’s low-cost labour model becomes an automation target rather than a permanent moat.

The Core Fallacy

The article assumes AI will make humans more productive inside the existing development model. Under the Discontinuity Thesis, that model fails when cognitive automation achieves durable cost and performance superiority, institutions cannot preserve human-only economic domains, and the majority lose access to economically necessary labour.

The World Bank’s 4.5 per cent automation-risk estimate is therefore treated as structural reassurance when it is only a snapshot of present occupational exposure. It does not measure cascading effects: fewer entry-level jobs, reduced wage bargaining power, substitution in export industries, intensified global competition, and the collapse of the human pipeline into skilled work.

“Retrain and redeploy” is the article’s ceremonial escape hatch. It assumes that new productive roles will appear at a scale comparable to the roles AI destroys. That assumption is the corpse hidden beneath the policy language. If AI is adopted precisely because it performs work with fewer people, retraining produces more qualified competitors for a shrinking labour market.

Hidden Assumptions

• Productivity gains will translate into broad employment rather than concentrated ownership gains.
• Bangladesh can preserve human labour markets through regulation or delayed adoption.
• New sectors will create enough jobs to absorb displaced garment and administrative workers.
• Entry-level work can be removed without destroying the pipeline that produces future expertise.
• Low wages will remain an advantage after automation lowers the value of low-cost labour.
• Better education and connectivity will create Sovereigns rather than merely more capable Servitors competing for fewer positions.
• The state has the institutional capacity to retrain workers, regulate adoption, and distribute gains.
• Growth will remain broadly consumable through wages rather than requiring transfers, dividends, or ownership claims.
• “Responsible use” can restrain competitive pressure once firms gain a cost advantage from replacing labour.
• The relevant timeline is gradual national development rather than rapid competitive displacement across borders.

Social Function

Classification: partial truth, transition management, and ideological anesthetic.

The article correctly identifies infrastructure, skills, and adoption speed as immediate constraints. It also correctly notices that entry-level jobs are disappearing before headline mass unemployment arrives. But it converts those facts into a familiar policy lullaby: educate the workforce, improve connectivity, regulate responsibly, and the human-centred economy will continue.

That prescription protects the legitimacy of existing institutions. It allows government and industry to describe the coming disruption as a skills problem instead of an ownership and participation crisis. The article acknowledges the machinery entering the factory, then asks the workers to become better-qualified applicants for the remaining seats.

The Verdict

This is a competent development memo built on an obsolete economic premise. AI may compress Bangladesh’s productivity gains into a decade, but it can also compress the timetable for Bangladesh’s labour-dependent growth model to become nonviable.

The article sees the first incision but misdiagnoses the patient. Bangladesh’s real vulnerability is not merely inadequate skills or delayed adoption. It is dependence on mass wage participation while AI progressively severs the labour-to-income-to-consumption circuit. Education, electricity, connectivity, and regulation are necessary transition infrastructure. They are not a survival strategy.

Under the Discontinuity Thesis, Bangladesh remains viable only by producing Sovereigns who control AI-linked capital, Servitors who remain indispensable to those Sovereigns, or institutions capable of securing broad ownership and distribution of automated productive capacity. Otherwise, “AI for development” becomes a more efficient way to eliminate the developmental role of Bangladeshi labour.