On 21 October 2026, Communications and Digital Minister Meutya Hafid said in a video on Komdigi's official account that there had been "no content takedown or takedown request" from the ministry over forest-and-land-fire (karhutla) content, according to Katadata's report. She said Komdigi is instead pushing platforms for algorithmic accountability and transparency. The denial followed TikTok's removal of a video by Siak Regent Afni Zulkifli, filmed after she planted trees on burned land. It had about 5 million views. Screenshots showed TikTok citing its Community Guidelines, and the video was restored after her appeal.
This is a content-moderation story, not a hate-speech story. Reporting does not show that any hate-speech or disinformation order was issued against karhutla material. But the episode lands on top of an enforcement architecture that makes the denial hard to evaluate.
The strongest case for the minister's position
The government has a real problem. Fires generate fast-moving falsehoods. On 17 September 2026 Komdigi officials listed the patterns they were seeing: fictional events, old footage presented as current, wrong locations, accurate images with misleading captions, and unverified accusations about who caused the fires, per InfoPublik's account of the briefing. Notably, that briefing described monitoring, counter-narratives and public education. It did not mention takedowns. A government that answers hoaxes with corrections rather than removals is doing what free-speech advocates ask. Taken at face value, the minister's account is consistent with that record.
Why the denial settles less than it should
The difficulty is that Komdigi also holds a much broader tool. Ministerial Decree 127/2026 covers electronic information containing disinformation and hate speech and classifies it as content that "troubles the public and disrupts public order." Kontan reported on 5 April 2026 that private user-generated-content platforms must act on a ministerial takedown order no later than four hours after receiving it, and that compliance will be tracked through the SAMAN moderation compliance system. The press-freedom group AJI says the decree was signed on 13 March 2026, calls its wording a "rubber article," and notes it involves no independent oversight or Press Council role, according to Republika. AJI tied it to the restriction of Magdalene's Instagram reporting on an acid attack against a KontraS activist.
I could not retrieve the decree's operative text, so I cannot say what procedural safeguards, notice rules or appeal rights it contains. That gap is itself the point. When the rule is vague, the clock is four hours, and orders are not published, a denial from the minister and a Community Guidelines notice from the platform can both be true, and the user still cannot tell which rule removed the post.
The karhutla complaints went beyond TikTok. Katadata reports that comedian Wendi Cagur and singer Andien suspected reduced reach, and influencer Tasya Farasya said her Instagram Story views fell from over a million to about 85,000 after she discussed the fires on 16 October. These are self-reported and unverified. Some users also circulated a Meta AI chatbot answer claiming a block at Komdigi's request, which is an unverified user claim and not evidence of anything. SAFEnet's Nenden Sekar Arum told Katadata the group was still reviewing reports and could not yet say whether platform rules or government requests were the cause.
The accountability gap is symmetrical
Komdigi is right that platform algorithms are opaque. Users cannot tell whether a drop in reach is a shadowban, a ranking change or a quiet legal request. But the same standard applies to the state. A ministry that demands transparency from platforms while operating a four-hour takedown channel with no public order log is asking for a trust it has not earned. Komdigi has shown it can be aggressive when it chooses. On 27 August, during protests at the parliament complex, it announced tighter monitoring of posts and live streams to prevent disinformation, hate speech and provocation, and promised firm action "in line with its authority," per InfoPublik's release.
A deadline this short also pushes platforms toward over-removal. A platform facing sanctions for a missed four-hour window will take down first and review later, and it will tend to enforce its own broad guidelines pre-emptively too. That makes it hard to separate state-directed removal from platform caution. The Siak video, which was restored on appeal, shows the cost: a regent's environmental advocacy, which the government says it welcomes, was removed by an automated or semi-automated system and fixed only after it went viral.
What proportionate regulation would look like
Three steps would let Komdigi's denial be tested rather than merely believed.
- Publish takedown orders. Release a periodic log of orders issued under Decree 127/2026, by category and platform, with the legal basis for each. If no karhutla orders exist, the log proves it.
- Lengthen the clock for speech. Four hours may suit child-safety content. For disinformation and "public order" categories, where context decides legality, review by someone outside the ministry should come before removal, or at least promptly after.
- Require platform notices to say who asked. When a post is removed or restricted, the notice should state whether the cause was a government request, a legal order or the platform's own policy. This is the transparency Komdigi asks of algorithms, applied to enforcement.
The minister's call for algorithmic accountability is a sound one, and it is more credible if the ministry accepts the same accountability for its own orders. Until it does, every unexplained disappearance of a post will be read as state action, whether or not it is.