Bangladesh Needs a Digital Shield, Not a Digital Gag

Social media has transformed the political battlefield. Facebook, YouTube, TikTok, X and messaging services such as WhatsApp are no longer simply channels of communication. They are places where political narratives are created, contested and amplified–often at a speed with which governments, journalists and traditional media struggle to keep pace.

That creates a difficult challenge for any democracy. An elected government has a legitimate responsibility to protect citizens and state institutions from foreign interference, terrorist propaganda, incitement to violence, coordinated disinformation and digitally organised attacks. But that responsibility cannot become a licence to suppress criticism, journalism or political opposition.

Bangladesh has particular reason to be cautious. We have already seen how broadly worded digital-security laws can become instruments of repression. The cases of writer Mushtaq Ahmed and cartoonist Ahmed Kabir Kishore became emblematic of concerns over the use of the Digital Security Act and its implications for freedom of expression. International rights groups also criticised the subsequent Cyber Security Act on similar grounds.

The country subsequently repealed the Cyber Security Act, 2023 and replaced it with the Cyber Security Ordinance, 2025. The ordinance, gazetted on 21 May 2026, removed nine controversial provisions of the earlier law.

Bangladesh now has the Cyber Security Act, 2026, while the government is considering further amendments to the law. Parliament also passed a Cyber Security (Amendment) Bill in June, including provisions concerning harmful and defamatory AI-generated content.

This is therefore a good moment to look beyond our borders. India, in particular, offers a model worth studying–not necessarily copying wholesale.

India's important lesson

India has taken a more assertive approach to intermediary responsibility through its Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021. The rules impose due-diligence obligations on online intermediaries, with additional requirements for significant social-media platforms.

Their February 2026 amendments are particularly relevant to Bangladesh because they introduce a specific framework for synthetically generated information–including AI-generated or manipulated audio, visual and audiovisual material. Platforms are required to employ technical measures concerning specified unlawful synthetic content, while other synthetic material covered by the rules is subject to disclosure and related safeguards.

The amendments also reduced, in specified circumstances, the period for intermediaries to act on government or court notices concerning unlawful content from 36 hours to three hours. The amended rules took effect on 20 February.

That is a powerful regulatory tool, and it is not without controversy. Three hours can be a remarkably short period in which to determine whether politically sensitive material is genuinely unlawful or merely offensive, satirical or critical.

Yet the underlying principle deserves Bangladesh's attention: social-media platforms cannot always be treated as passive noticeboards when their algorithms and recommendation systems can amplify information to millions of people almost instantaneously.

The threat is not criticism. It is manipulation.

Read More

Protecting Children: Understanding Causes of Child Abuse

Bangladesh should be careful about how it defines the threat.

A citizen criticising the government is not an enemy of the state. Peaceful protest, investigative journalism, satire and opposition politics are indispensable to a functioning democracy.

The greater danger is coordinated malicious activity.

A sophisticated influence operation can combine fake accounts, automated amplification, manipulated videos, fabricated documents, paid advertising and coordinated activity across several platforms. What appears to be spontaneous public sentiment may sometimes be the product of organised amplification.

Artificial intelligence makes the problem more acute. A fabricated video can make a political leader appear to say something he never said. A cloned voice can create a fictitious telephone conversation. A manipulated photograph can be circulated during a communal confrontation before anyone has established its authenticity.

The question, therefore, should not simply be: Who posted this?

It should also be: How was it amplified? Who coordinated the amplification? Was the network authentic? And did the platform take reasonable steps to prevent a demonstrably malicious operation from reaching millions?

That is a more sophisticated approach to digital security.

Regulate platforms, not political beliefs

The European Union offers another useful lesson.

Under the Digital Services Act, very large online platforms and search engines must assess and mitigate systemic risks, including risks to electoral processes, public discourse and fundamental rights. The European Commission's latest assessment continues to examine risks arising from the design and operation of major platforms.

Bangladesh need not reproduce the European system. But it could require major platforms operating here to assess risks associated with coordinated inauthentic behaviour, bot networks, election manipulation, deepfakes and cross-platform campaigns intended to provoke violence.

Australia offers another useful example. Its online-safety framework places obligations on digital services concerning recommender systems, anonymous accounts, generative AI and coordinated high-volume attacks across platforms.

The lesson from these jurisdictions is increasingly clear: platform accountability should focus not only on individual pieces of content but also on the systems that amplify harmful activity.

Speed matters–but so does due process

Digital misinformation has one particularly dangerous characteristic: speed.

During an election, communal confrontation or national emergency, a false claim can spread across several platforms before an official correction reaches the same audience.

India's amended framework recognises that reality. Bangladesh should consider whether it needs an emergency mechanism for narrowly defined circumstances–for example, credible incitement to imminent violence, terrorist propaganda, fabricated emergency instructions, impersonation of public institutions or synthetic material demonstrably intended to trigger serious disorder.

But speed cannot come at the expense of due process.

Government requests should identify the legal basis and the specific material concerned. Platforms should have a meaningful avenue of rapid review. Emergency decisions should be recorded and, where appropriate, subject to independent or judicial scrutiny.

The principle should be speed with accountability, not speed without accountability.

Bangladesh must not repeat its own mistakes

There is an equally important lesson at home.

The repeal of the 2023 law was significant precisely because nine controversial provisions were removed from the framework. The change reflected long-standing concerns about the criminalisation of speech and the potential misuse of digital-security legislation.

That lesson should not be lost as Bangladesh moves towards stronger regulation of AI-generated content.

The country should seek greater leverage over global technology companies without giving the executive an unrestricted power to determine what constitutes acceptable political speech.

Nor should allegations of foreign interference or “anti-state” activity be enough by themselves. Such allegations require evidence. Any official designation of a campaign as foreign-backed or coordinated should be based on verifiable evidence and subject to independent scrutiny.

The law should focus on conduct and demonstrable harm: covert foreign influence operations where supported by evidence, terrorist recruitment, incitement to imminent violence, cyberattacks, fraudulent impersonation, coordinated inauthentic networks and deliberate manipulation capable of producing serious harm.

It should not criminalise political disagreement.

What Bangladesh can borrow

Dhaka can selectively draw from several democratic models.

From India, it can learn about stronger intermediary obligations, rapid responses to unlawful content and explicit regulation of synthetic media.

From the European Union, it can consider systemic-risk assessments and greater transparency from the largest platforms.

From Australia, it can examine the advantages of a specialist regulator with powers to demand information and assess online-safety practices.

From the United Kingdom, it can consider an approach that links certain online offences to identifiable harm and intent rather than simply criminalising false information.

And from the United States, Bangladesh should take the enduring lesson that governments must exercise particular restraint when regulating political expression.

The objective should be neither unrestricted platforms nor unrestricted government power.

It should be accountable platforms operating within a rights-respecting legal framework.

Protect democracy, not the government from criticism

The distinction is fundamental.

The purpose of digital-security legislation should not be to protect a government from criticism. An elected government must be able to withstand criticism; that is the price of democratic accountability.

But neither should the state be expected to stand helplessly while networks fabricate statements, impersonate public officials, coordinate digital attacks or deliberately manufacture communal tensions.

The state has a legitimate duty to protect citizens and democratic institutions from such operations.

What Bangladesh needs is an independent, evidence-based mechanism capable of distinguishing a genuine security threat from political dissent.

One possibility would be a specialised digital-threat response mechanism involving cybersecurity professionals, law-enforcement agencies, election authorities and independent legal experts. Its mandate should be narrow. Every intervention should have a recorded legal basis and an auditable trail. Platforms should have avenues of appeal, while an independent oversight body should publish periodic data on government requests, removals and successful challenges.

Such safeguards would make it harder for digital-security powers to quietly become instruments of political censorship.

The real battle is over trust

The next generation of influence operations may not seek to persuade citizens to support one particular party. A more damaging objective could be to convince citizens that nothing is trustworthy–not the government, not the media, not election authorities and not even the results of an election.

That is a serious democratic vulnerability.

Bangladesh therefore needs to treat information integrity as part of national security. But information integrity cannot mean government-controlled truth.

The lesson from India's evolving framework is that powerful platforms must accept greater responsibility for the information ecosystems they create and amplify. The lesson from Europe, Australia and elsewhere is that such responsibility must be accompanied by transparency, proportionality and protection for legitimate speech.

Bangladesh should build a digital shield, not a digital gag.

A strong democracy should not fear criticism.

It should be equipped to resist manipulation.