JAKARTA: Two recent headlines from Indonesia’s digital landscape reveal a paradox in how online speech is policed. Media watchdogs Remotivi and the South-East Asia Freedom of Expression Network (SAFEnet) documented more than 100 social media posts critical of the government that were removed or restricted, nearly three-quarters of them traced to official requests from the Communication and Digital Affairs Ministry.
Meanwhile, Google’s latest Influence Operations Bulletin disclosed that it terminated more than 19,300 YouTube channels in the third quarter of 2026, and nearly 24,000 over the first nine months, for running coordinated campaigns promoting pro-government narratives in Indonesia.
The two developments seem to pull in opposite directions: one points to state suppression of dissent, the other to a platform crackdown on state-aligned propaganda. Yet both expose the same systemic failure: a profound lack of transparency in content moderation by governments and platforms alike.
The ministry’s response to Remotivi and SAFEnet’s findings was defensive rather than constructive. The two groups challenged the ministry to an open public debate over the 94 takedown requests attributed to it. Officials instead deflected, saying they lacked matching data and insisting the state is not a “censor”.
Refusing evidence-based dialogue is deeply troubling in a democracy. A ministry that can restrict public discourse, and that imposes tight turnaround windows for removals, is obliged to account for every request.
By sidestepping a public reckoning with civil society, the ministry erodes public trust and feeds suspicion that regulatory powers are being used to shield the powerful from scrutiny rather than serve the public interest.
Platforms deserve criticism for their own opacity. Google and Meta merit credit for publishing periodic transparency reports on enforcement actions and, in Meta’s case, government removal requests.
Google’s disclosure of its sweep against influence networks offers a valuable window into digital manipulation. But periodic aggregate statistics are not enough.
When platforms geo-block content or quietly remove posts without telling users the legal basis or the requesting authority, they become silent enforcers of state censorship.
The scale of Google’s takedowns also exposes the industrial-scale use of “buzzers” (Indonesia’s term for paid or coordinated social media accounts) and astroturfing networks in Indonesia’s public sphere.
According to Google’s findings, these networks did not merely cheer for officials; they also harassed citizens who criticized the government.
Google ultimately removed the channels, but the episode raises a pointed question: Why were such large networks able to flourish in the first place?
That question returns us to a long-standing demand from civil society and researchers: platforms must open their algorithms to scrutiny. Every major platform guards its recommendation and amplification systems as trade secrets, yet those systems routinely reward coordinated outrage, amplify buzzer-driven harassment and distort public debate long before any transparency report appears.
Without meaningful algorithmic transparency and independent auditing, aggregate takedown figures are little more than post-mortems of damage already done. The result is a dual failure.
Platforms are chronically slow to remove genuinely harmful content, such as coordinated harassment, doxing and state-sponsored disinformation, yet too quick to restrict legitimate political expression, often in quiet compliance with local laws or overbroad government requests.
Demanding accountability is not an argument against moderation. The goal is not an unmoderated digital Wild West but effective moderation that robustly defends free speech while protecting users from real-world harm. The two aims are not mutually exclusive.
Advances in artificial intelligence give platforms more sophisticated contextual tools than ever to distinguish authentic criticism, satire and investigative journalism from coordinated inauthentic behaviour and targeted abuse.
Getting this right is no longer optional, definitely not with the growing use artificial intelligence. Generative AI will sharply increase the volume and sophistication of synthetic propaganda, automated harassment and deepfakes, making moderation far harder.
If governments keep demanding takedowns behind closed doors, and platforms keep hiding behind opaque algorithms and partial disclosures, public discourse will be squeezed from both ends.
Only radical transparency, from regulators and tech giants alike, can keep the digital public square both safe and free. - The Jakarta Post/ANN
