Monsoon LedgerIndonesia AI governance

Why this exists

About

Monsoon Ledger is written like a small, open research project. This page sets it out the way a paper would: the background, the problem, what the ledger tries to do, its limits, and who is behind it.

Abstract

Indonesia has been announcing binding AI rules since January 2025. Monsoon Ledger follows those announcements in the language they are made in and checks each one against the record. It holds 6 sourced entries, 7 tracked instruments and 7 public promises. Of the 5 that set a deadline, 4 were missed and 1 is still open. 71 percent of the sources behind the entries were published in Bahasa Indonesia. Every claim is graded by the strength of its evidence, compared with the EU AI Act and frontier AI safety practice, and published as open data.

Entries
6
Dated promises
5
Dated promises missed
4
Sources in Bahasa Indonesia
71%

KeywordsAI governanceIndonesiapolicy trackingdocument analysisEU AI Actopen data

Background

Indonesia is writing its AI rules right now. The announcements, deadlines and consultations happen in Bahasa Indonesia, and little of that detail reaches readers outside the country.

The first public deadline in the ledger dates from 2025-01-13, when Meutya Hafid, Minister promised "a binding rule within three months". Two presidential regulations, a national AI roadmap and an AI ethics rule, have been described as nearly finished many times since. Meanwhile the only AI-specific instrument in force is a 2023 ministerial circular letter, which is guidance, not law.

Figure 1Where each tracked instrument stands, from announced to in force.
  1. 1Announced
  2. 2Draft
  3. 1Signed
  4. 3In force

Binding, or binding once in forceNot binding

The problem

From outside, an announced rule, a drafted rule, a signed rule and a rule in force can look like the same thing. Often the difference sits in a single Indonesian word in a ministry statement or a news report. Without a record that dates every promise and links every source, it is hard to tell how far Indonesia's AI governance has actually moved, and easy to report a promise as if it were law.

Figure 2The language gap. Sources behind the entries, by the language they were published in.

Bahasa Indonesia 10English 4

Objectives

  1. RecordEvery public promise about Indonesia’s AI rules, with the date, the person who made it and a link to the report.
  2. GradeEvery claim by the strength of its evidence: the legal text itself, an official statement, or a press report.
  3. CompareEach instrument with the EU AI Act and with frontier AI safety practice, so readers can see what is and is not covered.
  4. OpenAll of it as open data and open code, so anyone can check the reading, reuse it or correct it.

Scope and limits

  • Indonesia only, until the Indonesian ledger has twelve entries. Other ASEAN countries come after that. The Radar already watches them.
  • Data protection appears only where it shapes AI governance, such as breach reporting and the regulator.
  • Desk research only. No interviews and no information requests so far. Every piece of evidence is published and linkable.
  • An independent project, not legal advice. The method and its limits are set out on the Method page.

Contribution

The ledger keeps the original Indonesian wording next to an English reading, so a reader abroad can see exactly what was said. It shows the evidence level on every claim, keeps a public log of corrections, and connects entries, instruments, promises, officials and documents in one ontology.

Two parts are meant to be reused. The promise format is an open, documented way to record government promises with deadlines derived by published rules and outcomes computed rather than judged, with a schema, worked examples and a kit for a ledger in another country. Every promise also has a permanent receipt: what was said, by whom, when it fell due and how late it is, with the source, a preserved copy, citation formats and a card that stays accurate when shared.

A scanner checks the news twice a week, but only the author can publish a claim. Claude, an AI model, helps build the site, draft and translate text for the author to review, and give an optional second opinion on scanner headlines; how it is used is set out on the Technical system page, together with the whole build history.

About the author

I'm Stephen Behren Tanjung, an Indonesian living in Deventer, the Netherlands. I read Bahasa Indonesia natively and I live under EU rules, so I read Indonesian AI policy in the language it is announced in and set it beside the rules I live with.

I hold a Bachelor of Arts in Creative Business from Hogeschool Utrecht (2020 to 2024), with a minor in Big Data and Design. The programme covered data storytelling, human-centred design, and AI and creativity. In 2022 my team won first place at the Big Data and Design Hackathon at Hogeschool Utrecht, redesigning the Water Point Data Exchange platform with Akvo so that people deciding on rural water access could find the data faster.

My work sits where visual design, multimedia and data meet. In 2025 I was a multimedia content creator at ACAPS in The Hague, making graphics, short videos and infographics that reached donors and humanitarian responders. Since January 2026 I have been a content marketing specialist at Rush Send. Before that I edited video as a volunteer in Sukabumi, Indonesia, from 2015 to 2019, and made content for the Indonesian Student Association in Deventer.

In 2026 I completed a certificate course in Dutch law at the University of Groningen and Anthropic's AI Fluency course. That is the background I bring to reading legal texts and AI policy. Monsoon Ledger is an independent project. It is not legal advice.

ReadsIndonesiaPolicy in Bahasa Indonesia
Lives underEU rulesDeventer, the Netherlands
  • Indonesiannative
  • Englishfull professional
  • Dutchconversational
  1. 2015 to 2019WorkVideo editor, volunteer, Mawar Sharon, Sukabumi, Indonesia
  2. 2019 to 2021WorkContent creator, volunteer, Indonesian Student Association, Deventer
  3. 2020 to 2024StudyBA Creative Business, Hogeschool Utrecht, with a minor in Big Data and Design
  4. 2022AwardFirst place, Big Data and Design Hackathon, with Akvo and the Water Point Data Exchange
  5. 2023WorkSocial media designer and content creator, internship, THYSOL Group, Enschede
  6. 2025WorkMultimedia content creator, ACAPS, The Hague
  7. 2026WorkContent marketing specialist, Rush Send
  8. 2026StudyCertificates in Dutch law (University of Groningen) and AI Fluency (Anthropic)
  9. 2026NowStarts Monsoon Ledger, first public entries on 22 October

Portfolio LinkedIn

Why Indonesian first

Policy is announced in Bahasa Indonesia, and the exact words matter. A draft, a signature and a rule in force are different things, and the difference often sits in a single word. So key quotes stay in the original, with an English line under them, and every source is linked so you can check my reading.

Why the name

The monsoon is forecast and announced long before it arrives. Indonesia's AI rules have worked the same way: announced, given a date, and announced again. The ledger records what was promised and what actually showed up.

Where this is going

Monsoon Ledger started as a way to check what governments promise about AI against what happens. The longer question underneath it is what a country needs in place to notice, report and respond when advanced AI systems fail. I want to work on that full time, on AI safety and governance, and I am building the evidence in public as I go: this ledger, the analysis papers, the incident reference and the reading list. Corrections and criticism make the work better, and are welcome.