Dina Mahmud
DTech Bangladesh
When Bill Gates told Reuters recently that no government on Earth is genuinely prepared for what artificial intelligence is poised to unleash on society, he wasn’t singling out any one nation for criticism. He was pointing to a structural truth that technologists in this space have long recognized: policy has spent years racing to keep pace with capability, and it hasn’t caught up yet.
Bangladesh offers a useful lens for understanding why that gap matters, not because the country lags unusually far behind, but because it illustrates precisely what genuine preparedness demands beyond good intentions on paper. This article aims to open that conversation.
Where Bangladesh Actually Stands
To Bangladesh’s credit, this isn’t a first attempt at building something from nothing. A National AI Policy covering 2026 to 2030 has already passed through public consultation, introducing a risk-based regulatory framework, limits on mass surveillance and social scoring, and a proposed National AI Act intended to establish a dedicated regulatory authority by 2027. That marks genuine progress compared to the country’s earlier national AI strategy from 2019 and 2020, an initiative that generated thorough roadmaps but left little lasting impact. That history deserves attention. As a technologist, I’ve come to distrust policy documents that sound convincing, and I’m even more skeptical of the idea that writing a strategy is equivalent to building real capacity.
Bangladesh currently holds the 75th spot on the Oxford Insights Government AI Readiness Index, a middling position that reflects both meaningful institutional effort and genuine structural weaknesses. A recent readiness assessment conducted with international partners was blunt about what those weaknesses involve: fragmented data systems, a shortage of the computing hardware essential to modern AI, academic curricula that have fallen behind, and AI ethics education that is nearly nonexistent within the system. None of these amounts to abstract policy failure. These are the tangible reasons a country can have an impressive draft law in place and still be unready for the point when AI systems begin influencing decisions about people’s jobs, healthcare, and access to public services.
The Human Cost Hiding Inside the Technical Gaps
Here is where the tone needs to shift from infrastructure audit to something closer to a warning about human value. Bangladesh has built one of the largest freelance digital workforces in the world, hundreds of thousands of people earning meaningful foreign income through digital and IT enabled services. Every one of these gaps, the GPU scarcity, the missing ethics curriculum, the fragmented data protection rules, translates directly into risk for that workforce and for ordinary citizens who will increasingly interact with automated systems in banking, healthcare, and government services without adequate protection or recourse. A country can have excellent policy language about human centric innovation and still leave its most vulnerable digital workers and rural populations exposed if implementation lags ambition, as it did the last time this exercise was attempted.
Social harmony depends on people trusting that new systems will not quietly widen inequality or replace their livelihoods without a plan for what comes next. That trust is not won through white papers. It is won through visible, functioning safeguards.
Closing the Gaps Without Losing Sight of People
Three concrete steps need to happen right away, and none of them require anything exotic. Computing infrastructure can no longer be aspirational language tucked into a policy draft; shared national compute capacity needs real procurement timelines and budget commitments behind it, not just a statement of intent. AI ethics and literacy need to enter school curricula now, not once the National AI Act passes, since the workforce that will eventually implement and audit these systems is being educated at this very moment.
Most critically, any rollout of AI into public services, whether that’s health records or social safety net disbursement, needs a human appeals channel built in from the outset, so no citizen ends up arguing with an algorithm with no way to reach an actual person. These aren’t distant priorities to revisit later; they’re the groundwork that determines whether everything else in the policy functions.
Gates likened AI’s arrival to a disaster film, with humanity forced to unite against something truly alien. The comparison is dramatic, but it lands on something real: readiness isn’t a document sitting in a drawer, it’s a coordinated, funded, continuously tested capability. Bangladesh has finally written a stronger script than it managed in 2019. Whether it follows through with the actual work this time will decide whether AI becomes the tool that narrows the country’s inequality, as its policymakers hope, or simply one more technology that outpaces the institutions meant to protect ordinary people.
