AI Governance Needs the Media Too
- AI governance needs journalism
- Verification and accountability are crucial
- Bangladesh needs inclusive AI dialogue
Artificial intelligence is rapidly becoming a new infrastructure of economic, technological and political power. Countries are competing for semiconductors, cloud infrastructure, data centres, advanced computing capacity and skilled human resources. At the same time, governments and international institutions are considering how to keep powerful AI systems safe, fair and accountable.
For Bangladesh, this transformation presents both opportunities and risks.
We must think about computing capacity, data governance, Bangla-language AI, cybersecurity, human resources and Bangladesh’s position in global AI governance. But another important question is often overlooked: What will be the role of the media?
AI governance cannot be left solely to technologists, government ministries or diplomats. Citizens need to understand how AI is changing their jobs, education, agriculture, banking, public services, elections, information flows and everyday lives.
The media, therefore, should be viewed not merely as an observer of the AI transformation, but as part of the democratic infrastructure of AI governance.
Sovereignty Is More Than GPUs
Discussions about AI sovereignty usually focus on questions such as who controls the chips, where data centres are located, who provides cloud services and whether developing countries have sufficient computing capacity.
These are important questions. But sovereignty cannot be measured only in GPUs, servers and megawatts.
If Bangladesh uses technology without having the independent capacity to understand what that technology does, whose interests it serves, what data it uses and who is accountable when something goes wrong, then its sovereignty remains incomplete.
This is where journalism has an important role.
The media must ask: Who owns the AI system being used by a government agency? Where is government data being stored? Is sensitive information being transferred outside the country? Who is testing the algorithms? If a citizen is harmed by an automated decision, where can they seek redress?
These are not merely technology questions. They are questions of power, rights and accountability.
Translating Technology Into People’s Language
Most ordinary people do not understand terms such as compute sovereignty, foundation models, algorithmic accountability or data governance.
The media's job is to turn these concepts into questions about people's lives:
- How will AI change the work of garment workers?
- Will farmers receive accurate market and weather information?
- Will government services become easier?
- Will surveillance increase?
- Will Bangla-speaking people receive AI services comparable to those available in English?
- Which jobs could disappear?
- Which new jobs could emerge?
These questions make the discussion about technology accessible to everyone.
Neither Hype nor Fear
AI coverage often falls into two extremes. On one side, there is enthusiasm—as if AI will automatically guarantee progress. On the other, there is fear—as if AI will inevitably eliminate jobs, journalism and human decision-making.
Both approaches are misleading.
The impact of AI depends on how it is developed, used and governed.
If a government agency claims that AI has improved efficiency, the media should ask: What is the evidence? If a company claims that its algorithm is unbiased, journalists should ask how the company tested it.
The fundamental journalistic question remains the same:
Who holds the power, and how is that power being used?
Deepfakes Will Test Journalism
Generative AI can now produce realistic fake images, voices, videos and written content. During elections, political tensions, communal conflicts or disasters, the impact of AI-generated misinformation could be devastating.
Fake audio can influence public opinion. Fake images can provoke violence. Deepfakes can be weaponised against women journalists and human-rights activists.
Therefore, verification cannot remain the responsibility of fact-checkers alone. Every newsroom must treat verification as a fundamental part of journalism.
Bangladesh’s media outlets need stronger capabilities in digital investigation, source verification, identifying AI-generated content and developing transparent publication policies.
In the AI era, the demand for speed will increase. So will the value of verification.
Take AI Discussions Beyond Dhaka
If Bangladesh's AI conversation remains concentrated in Dhaka, conducted primarily in English and dependent largely on technology experts, a significant part of the country will be left out.
Women face emerging risks from technology-enabled gender-based violence and deepfakes. Rural communities could benefit from AI in agriculture and healthcare, but internet connectivity, cost and language remain major barriers. People with disabilities could gain significantly from assistive AI—provided that AI systems are designed inclusively.
Community and local media therefore have an important role. They can explain AI in simple Bangla, bring local experiences into national policy discussions, and communicate people's concerns and expectations.
Democratising AI means including the people whose lives will be transformed by the technology in the conversation about it.
Bangla-Language AI Is a Public-Interest Issue
Many AI systems still perform better in English than in Bangla. If this disparity persists, a new form of digital inequality could emerge.
If a government chatbot provides accurate answers in English but performs poorly in Bangla, the problem is not merely a technical flaw. It becomes an issue of equal rights and access.
The media should ask:
- Who is creating Bangla-language datasets?
- Do those datasets represent regional dialects and accents?
- Who owns Bangla-language data?
- Do Bangladeshi researchers have access to the computing capacity they need?
Language is therefore not only a matter of culture. It is also an issue of economic and technological self-determination.
Newsrooms Need Their Own AI Policies
The media should question how others use AI, but news organisations must also take responsibility for their own use of the technology.
AI can assist journalists with transcription, translation, research, data analysis and document review. But what happens if confidential-source information is uploaded to a commercial AI platform? What if AI generates false information? When should readers be told that AI was used in producing content?
Every responsible media organisation should have a Responsible AI Policy covering source confidentiality, fact-checking, copyright, data protection, transparency and human editorial responsibility.
The principle should be simple:
AI should assist journalism, not replace journalistic responsibility.
A Bangladesh Media and AI Governance Initiative
Bangladesh could establish a Bangladesh Media and AI Governance Initiative, bringing together national media organisations, community media, journalism education institutions, fact-checkers, civil society, universities, technologists and relevant government agencies.
Such an initiative could:
- Increase AI literacy among journalists;
- Build deepfake-verification capabilities;
- Develop newsroom AI guidelines;
- Produce Bangla-language resources on AI journalism;
- Strengthen investigative journalism on government AI procurement, algorithms and global AI governance.
This would help improve the quality and accountability of journalism in the AI era.
Conclusion
Bangladesh’s AI future cannot be measured only by the number of data centres, GPUs purchased or AI companies established.
The real test lies elsewhere.
Do citizens understand how the technology is being used?
Can journalists question powerful AI systems?
Can people challenge automated decisions?
Are women, young people, rural communities and marginalised groups participating in policy discussions?
Computing power is necessary. Data is necessary. Infrastructure is necessary. But democratic oversight is equally necessary.
In the AI era, the media’s role is not merely to report on the technological future after it arrives.
A new responsibility of journalism is emerging: to help society understand, shape and actively participate in Bangladesh’s technological future.
Author: Chief Executive Officer, Bangladesh NGOs Network for Radio and Communication (BNNRC).
Note: The full opinion published in the Opinion section represents solely the author’s views. The Digital Bangla Media authorities have no association with the views expressed. The article is published without editorial alteration as an expression of diverse opinions in the media. Any offence or agitation caused by the article is entirely a personal matter for the individual concerned.
Artificial intelligence is rapidly becoming a new infrastructure of economic, technological and political power. Countries are competing for semiconductors, cloud infrastructure, data centres, advanced computing capacity and skilled human resources. At the same time, governments and international institutions are considering how to keep powerful AI systems safe, fair and accountable.
For Bangladesh, this transformation presents both opportunities and risks.
We must think about computing capacity, data governance, Bangla-language AI, cybersecurity, human resources and Bangladesh’s position in global AI governance. But another important question is often overlooked: What will be the role of the media?
AI governance cannot be left solely to technologists, government ministries or diplomats. Citizens need to understand how AI is changing their jobs, education, agriculture, banking, public services, elections, information flows and everyday lives.
The media, therefore, should be viewed not merely as an observer of the AI transformation, but as part of the democratic infrastructure of AI governance.
Sovereignty Is More Than GPUs
Discussions about AI sovereignty usually focus on questions such as who controls the chips, where data centres are located, who provides cloud services and whether developing countries have sufficient computing capacity.
These are important questions. But sovereignty cannot be measured only in GPUs, servers and megawatts.
If Bangladesh uses technology without having the independent capacity to understand what that technology does, whose interests it serves, what data it uses and who is accountable when something goes wrong, then its sovereignty remains incomplete.
This is where journalism has an important role.
The media must ask: Who owns the AI system being used by a government agency? Where is government data being stored? Is sensitive information being transferred outside the country? Who is testing the algorithms? If a citizen is harmed by an automated decision, where can they seek redress?
These are not merely technology questions. They are questions of power, rights and accountability.
Translating Technology Into People’s Language
Most ordinary people do not understand terms such as compute sovereignty, foundation models, algorithmic accountability or data governance.
The media's job is to turn these concepts into questions about people's lives:
- How will AI change the work of garment workers?
- Will farmers receive accurate market and weather information?
- Will government services become easier?
- Will surveillance increase?
- Will Bangla-speaking people receive AI services comparable to those available in English?
- Which jobs could disappear?
- Which new jobs could emerge?
These questions make the discussion about technology accessible to everyone.
Neither Hype nor Fear
AI coverage often falls into two extremes. On one side, there is enthusiasm—as if AI will automatically guarantee progress. On the other, there is fear—as if AI will inevitably eliminate jobs, journalism and human decision-making.
Both approaches are misleading.
The impact of AI depends on how it is developed, used and governed.
If a government agency claims that AI has improved efficiency, the media should ask: What is the evidence? If a company claims that its algorithm is unbiased, journalists should ask how the company tested it.
The fundamental journalistic question remains the same:
Who holds the power, and how is that power being used?
Deepfakes Will Test Journalism
Generative AI can now produce realistic fake images, voices, videos and written content. During elections, political tensions, communal conflicts or disasters, the impact of AI-generated misinformation could be devastating.
Fake audio can influence public opinion. Fake images can provoke violence. Deepfakes can be weaponised against women journalists and human-rights activists.
Therefore, verification cannot remain the responsibility of fact-checkers alone. Every newsroom must treat verification as a fundamental part of journalism.
Bangladesh’s media outlets need stronger capabilities in digital investigation, source verification, identifying AI-generated content and developing transparent publication policies.
In the AI era, the demand for speed will increase. So will the value of verification.
Take AI Discussions Beyond Dhaka
If Bangladesh's AI conversation remains concentrated in Dhaka, conducted primarily in English and dependent largely on technology experts, a significant part of the country will be left out.
Women face emerging risks from technology-enabled gender-based violence and deepfakes. Rural communities could benefit from AI in agriculture and healthcare, but internet connectivity, cost and language remain major barriers. People with disabilities could gain significantly from assistive AI—provided that AI systems are designed inclusively.
Community and local media therefore have an important role. They can explain AI in simple Bangla, bring local experiences into national policy discussions, and communicate people's concerns and expectations.
Democratising AI means including the people whose lives will be transformed by the technology in the conversation about it.
Bangla-Language AI Is a Public-Interest Issue
Many AI systems still perform better in English than in Bangla. If this disparity persists, a new form of digital inequality could emerge.
If a government chatbot provides accurate answers in English but performs poorly in Bangla, the problem is not merely a technical flaw. It becomes an issue of equal rights and access.
The media should ask:
- Who is creating Bangla-language datasets?
- Do those datasets represent regional dialects and accents?
- Who owns Bangla-language data?
- Do Bangladeshi researchers have access to the computing capacity they need?
Language is therefore not only a matter of culture. It is also an issue of economic and technological self-determination.
Newsrooms Need Their Own AI Policies
The media should question how others use AI, but news organisations must also take responsibility for their own use of the technology.
AI can assist journalists with transcription, translation, research, data analysis and document review. But what happens if confidential-source information is uploaded to a commercial AI platform? What if AI generates false information? When should readers be told that AI was used in producing content?
Every responsible media organisation should have a Responsible AI Policy covering source confidentiality, fact-checking, copyright, data protection, transparency and human editorial responsibility.
The principle should be simple:
AI should assist journalism, not replace journalistic responsibility.
A Bangladesh Media and AI Governance Initiative
Bangladesh could establish a Bangladesh Media and AI Governance Initiative, bringing together national media organisations, community media, journalism education institutions, fact-checkers, civil society, universities, technologists and relevant government agencies.
Such an initiative could:
- Increase AI literacy among journalists;
- Build deepfake-verification capabilities;
- Develop newsroom AI guidelines;
- Produce Bangla-language resources on AI journalism;
- Strengthen investigative journalism on government AI procurement, algorithms and global AI governance.
This would help improve the quality and accountability of journalism in the AI era.
Conclusion
Bangladesh’s AI future cannot be measured only by the number of data centres, GPUs purchased or AI companies established.
The real test lies elsewhere.
Do citizens understand how the technology is being used?
Can journalists question powerful AI systems?
Can people challenge automated decisions?
Are women, young people, rural communities and marginalised groups participating in policy discussions?
Computing power is necessary. Data is necessary. Infrastructure is necessary. But democratic oversight is equally necessary.
In the AI era, the media’s role is not merely to report on the technological future after it arrives.
A new responsibility of journalism is emerging: to help society understand, shape and actively participate in Bangladesh’s technological future.
Author: Chief Executive Officer, Bangladesh NGOs Network for Radio and Communication (BNNRC).
Note: The full opinion published in the Opinion section represents solely the author’s views. The Digital Bangla Media authorities have no association with the views expressed. The article is published without editorial alteration as an expression of diverse opinions in the media. Any offence or agitation caused by the article is entirely a personal matter for the individual concerned.
