AI Autonomy or Algorithmic Dependence: Bangladesh at a Critical Crossroads

The key question is whether, in the future, we will be able to retain the right to think in our own language, within our own realities, and through our own social consciousness. Bangladesh must decide now: will it remain merely a market for global AI, or will it become a partner that safeguards its knowledge base, language, and public interest? Time still remains—but not for long.

While wars in the Middle East, missiles, geopolitics, alliances, and energy politics dominate global discourse, another quieter conflict is rapidly intensifying beneath the surface—one that may have even deeper implications for countries like Bangladesh. This is the struggle over the future of human intelligence, knowledge, and agency. Recently, at BlackRock’s US Infrastructure Summit, OpenAI CEO Sam Altman described a future where AI could be consumed like electricity or water—metered and paid for as a service. This is not merely a business model proposition; it signals a world where the capacity to think itself may gradually become embedded within corporate infrastructure.

The question, then, is whether Bangladesh will view this shift solely as technological progress, or as a matter of knowledge, language, policy, and sovereignty. If AI becomes a service controlled by a handful of global platforms, developing countries may remain users, not creators; consumers, not regulators. Knowledge would no longer remain a freely accessible human resource but could become a rented capability. This concern is not abstract for Bangladesh. The country’s draft National Artificial Intelligence Policy 2026–2030 explicitly identifies “digital sovereignty” as a core principle, emphasizing the need for effective control over critical data, digital infrastructure, and AI systems.

This is where the issue becomes serious. Policymakers in Bangladesh are beginning to recognize that AI is not merely about code, chatbots, or automation—it represents a new configuration of power. The draft policy acknowledges linguistic and cultural diversity as central to the country’s digital future. It clearly states that Bangla and other nationally significant languages must be given due importance in AI models, datasets, and digital services to ensure inclusivity and relevance. The policy also proposes the development of a state-led national Bangla language model, underscoring that language in the AI era is not only a cultural concern but also a matter of power.

However, policy recognition and practical preparedness are not the same. A UNESCO-supported AI Readiness Assessment for Bangladesh shows that although mobile penetration is high, only 44.5 percent of the population uses the internet. In urban areas, internet usage stands at around 66.8 percent, while in rural areas it is just 29.7 percent. Among urban users, 71.3 percent are men and 62.4 percent are women; in rural areas, the figures drop to 36.6 percent for men and only 23 percent for women. The same report highlights that infrastructural weaknesses and frequent electricity issues remain major barriers in rural regions. In terms of computing capacity, Bangladesh ranks 73rd out of 76 countries in colocation data center capability. In other words, discussions about AI are taking place at a time when a large segment of the population still lacks basic digital equity.

In this context, the greatest risk is confusing “digital participation” with “cognitive autonomy.” Owning a smartphone does not equate to controlling knowledge. Using platforms like ChatGPT, Gemini, or Copilot does not necessarily mean participating in knowledge production. On the contrary, there is a risk that users may become dependent on models that do not adequately understand their language, society, legal realities, local economy, regional history, religious sensitivities, or rural experiences. In such cases, AI would not merely assist—it would begin to interpret reality on behalf of users. And a society that cannot interpret its own reality in its own language gradually becomes subject to external algorithmic perspectives.

This risk is particularly evident for Bangladesh. UNESCO assessments indicate that greater attention is needed for AI development in Bangla and minority languages. Due to the lack of open and diverse datasets, AI outputs in Bangla remain weak, and translations from English to Bangla often produce low-quality results. The implication is clear: if linguistic infrastructure remains underdeveloped, Bangla-speaking populations will remain users in the future knowledge economy, not standard-setters. In that case, mistranslations, cultural mismatches, information distortion, and policy misinterpretations will not remain technical issues—they will evolve into democratic and social challenges.

The education sector is not immune to these concerns. UNESCO data shows that only slightly more than half of lower secondary and secondary schools have access to the internet for educational purposes. Before envisioning an AI-driven education system or knowledge economy, it is essential to assess the strength of foundational infrastructure. If primary students, rural youth, women, marginalized communities, and Bangla-medium learners remain outside AI literacy, then AI will not only create new skills but also new inequalities. In the future labor market, those who understand algorithms will advance, while those who merely consume their outputs will fall behind.

Therefore, the real question before Bangladesh is not whether to adopt AI, but what kind of AI, under whose governance, in which language, and with what accountability. Two extremes must be avoided: blindly embracing technology as a solution to all problems, and rejecting AI out of fear. Bangladesh needs a third path—human-centered, Bangla-supported, transparent, accountable, and public-interest-driven AI development. The UNESCO readiness assessment also points in this direction, recommending the finalization of an inclusive national AI policy, strengthening data protection and cybersecurity frameworks, building datasets in Bangla and minority languages, increasing public awareness, promoting inclusive skill development—especially for women—and expanding investment in research and infrastructure.

The role of the state extends beyond legislation. Universities must advance research. Civil society must engage in discussions on algorithmic accountability. Bangla language technology must be prioritized at the national level. Transparency must be ensured in public procurement involving AI. In sectors such as the judiciary, administration, healthcare, education, agriculture advisory, and social protection, citizens must be able to distinguish between human decisions and machine recommendations, and understand avenues for appeal. Otherwise, “smart governance” could easily turn into “opaque governance,” leading to an erosion of human agency.

For Bangladesh, this struggle ultimately revolves around three core issues: language, autonomy, and equity. Language, because no AI future can be just if Bangla and local knowledge are excluded. Autonomy, because if thinking becomes dependent on external platforms, digital sovereignty will remain theoretical. Equity, because if the benefits of AI are concentrated among urban, English-speaking, technologically privileged groups, it will create a new architecture of inequality rather than development.

The wars in the Middle East will eventually reach some form of political resolution. But the struggle over control of human knowledge, language, and judgment will be longer, more subtle, and more decisive for countries like Bangladesh.

Because the question is not only who will build technology, but whether we will retain the right to think in our own language, within our own realities, and through our own social understanding. Bangladesh must decide now: will it remain merely a market for global AI, or will it become a partner safeguarding its own knowledge base, language, and public interest? Time still remains—but not for long.

Author: A H M Bazlur Rahman, Digital Democracy Specialist; Chief Executive Officer, Bangladesh NGOs Network for Radio and Communication (BNNRC); Ambassador for Responsible Artificial Intelligence and Antimicrobial Resistance (AMR) for Bangladesh

Note: All opinions expressed in the Op-Ed section are solely those of the author and do not reflect the views of Digital Bangla Media authorities. As a reflection of diverse perspectives, the article is published without editorial alteration. Any offense or dissatisfaction arising from it is entirely a personal matter.