Indonesia governs AI with a handful of binding rules resting on a much larger base of ethics circulars, sector codes, and drafts still working through the State Secretariat. This site maps those rules, watches whether the binding ones get followed, and tracks the court cases that turn on AI.
Instruments are stacked by how much force they carry, from binding law with sanctions down to voluntary regional principles. That ordering is the argument: almost everything still sits below the binding line. Each entry links to its source where we've verified one; the rest are on the way.
Tap a marker
Five are drafts in the pipeline, shown dashed; the rest are in force, adopted, or assessments. Among the binding rules, the binding and principles corner stays empty: where the law binds, it also prescribes. The draft Perpres are the first instruments pointing into it.
The Observatory monitors the AI-specific obligations in Indonesian law that are actually binding on operators. Today there is exactly one: Article 47 of Permendag 19/2026. That scarcity is the point. It is the first and so far only place where the law says shall rather than should.
What operators must do when they deploy AI. Three of the six are visible from outside the company: disclosure, internal governance, and a complaint channel. Those are the ones the Observatory checks below.
AI may be used, provided the use conforms to applicable law. Deployment now carries legal responsibility.
The operator is legally responsible for its AI; outcomes cannot be blamed on the system itself.
Must inform or label consumers when goods, content, recommendations or promotions are AI-driven. Core duty tracked here.
AI-mediated information must be truthful, clear, accurate and accountable.
Platforms must maintain AI governance proportionate to the risk profile of their deployment.
A mechanism to challenge AI outputs and receive a substantive response.
Three of Article 47's six duties can be checked from the outside: whether consumers are told AI is in use, whether the operator keeps an internal AI policy proportionate to its risk, and whether there is a channel to complain about AI-related harm. Each operator below is checked against those three. Entries record observed practice on a given date, not a verdict of compliance. Assessed companies are open to tap; the rest are marked Coming soon until they are reviewed.
Article 47 carries no direct sanction, so this monitor does not label any operator "non-compliant." It records what is publicly observable on a given date, with evidence, and lets readers draw their own conclusions.
This is an independent research project and does not constitute legal advice.
A running record of court cases and investigations in Indonesia where AI is part of the facts. Most of it is deepfake fraud and synthetic sexual content, now joined by a civil tort over an AI face-swap and a divorce ruling on how to authenticate digital evidence. Five rulings are decided; the rest sit at investigation or prosecution.
Tap a case to read the facts, the legal basis, and where it stands. Status reflects the most recent stage we have on record.
Indonesian courts are deciding these cases without any AI-specific law, leaning on the Criminal Code, the ITE Law, the PDP Law, the Civil Code, and, in the election case, constitutional principle. Most turn on the substance of synthetic media even though no settled way to authenticate it yet exists, though one recent ruling has begun to require digital forensics. As of June 2026 Indonesia still has no standalone deepfake or AI-content statute.
Compiled from a June 2026 review of public sources, with status noted per case.
Each entry links to a traceable source: an official decision where one exists, otherwise credible reporting. This is a record of reported proceedings, not legal advice.
DJKI holds the human line: purely AI-made works are not recorded, patent inventorship stays human, and the copyright bill codifies a human-contribution test rather than extending protection to machines.
The AI-IP Index reads a country on two questions that usually get folded together. Accommodation asks how far the law will go to protect intellectual property that a machine helped produce. Institution asks how much administrative machinery sits behind that law: its guidance, its international submissions, its disclosure duties.
A country can be generous on paper and have built almost nothing, or cautious and highly organised, so the two are scored apart and reported as a pair. Each rests on three questions, scored 0, 0.5 or 1.
How far the law will protect intellectual property that a machine helped make.
How much administrative machinery sits behind that law.
The scale runs the same way on both axes: 0 is the open or developed end, 1 is the restrictive or absent one, and 0.5 sits between for anything conditional or untested. Read as a pair, the two scores show whether AI-assisted work can be protected here, and whether the office has said so in writing.
Within each axis the three questions carry equal weight. That is a deliberate choice, not an oversight. An AI developer might reasonably value freedom to train above patent inventorship, but weighting is where indices lose their credibility, so the score stays a flat mean and the choice is stated in the open. Each cell also cites a statute section, a case or an office page, so any single score can be challenged without disputing the rest.
Two positions are worth naming. A country can be open on the law but thin on machinery, an open-but-undeveloped stance that bets on ambiguity. Or it can be restrictive on rights yet run an active office, a restrictive-but-developed stance that reflects a settled policy choice rather than a gap. The pair of scores tells those two apart where a single ranking would blur them.
A rubric is only as good as the results it produces on cases where the answer is already known. These three jurisdictions sit outside this series and are scored with the same six questions, as a check. Their law is the best documented anywhere, so if the scale puts them where the law actually sits, the scale is doing its job. They are reference points, not editions.
Firm judicial settlement at both poles and the weakest statutory position on training. The courts have closed the door on AI inventors and AI authors, while everything on training rides on unresolved fair-use litigation. There is no per se duty to declare AI use in a filing, only a candour duty that bites where the use is material.
Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022); Thaler v. Perlmutter (D.C. Cir. 2025), cert. denied 2 Mar 2026; Bartz v. Anthropic (N.D. Cal. 2025); USPTO revised inventorship guidance, 90 Fed. Reg. 54636 (28 Nov 2025); 89 Fed. Reg. 58128 (17 Jul 2024); 89 Fed. Reg. 25609 (11 Apr 2024).
The most permissive statutory position on training and the strongest transparency duty in force, sitting over an authorship question no court has answered. Read the label carefully: training and transparency run through EU law, but patents run through the EPO under the European Patent Convention, which is not an EU institution and covers 39 states.
Directive (EU) 2019/790, arts. 3 and 4; Regulation (EU) 2024/1689, art. 53(1)(c) and (d); EPO J 8/20 (2021); EPO Guidelines G-II 3.3.1 (2026 edition); Infopaq C-5/08; Painer C-145/10.
Liberal on patents since the Supreme Court abandoned the Aerotel test in February 2026, conservative on copyright, and currently thin on AI-specific machinery after the dedicated AI examination guidance was withdrawn. Section 9(3) nominally protects a work with no human author for 50 years, the widest AI-output right on paper, but no court has ever applied it to generative AI.
CDPA 1988 ss. 9(3), 12(7), 29A, 178; Thaler v Comptroller-General [2023] UKSC 49; Comptroller-General v Emotional Perception AI [2026] UKSC 3; UKIPO practice notice, 14 Jul 2026; Data (Use and Access) Act 2025 ss. 135 to 137.
The check earned its keep twice. The untested rule held: section 9(3) of the UK Copyright Act reads like the widest AI-output right anywhere, yet because no court has applied it to generative AI it scores as untested rather than open, which is the correct answer and the one a headline reading would miss. The check also exposed a limit worth stating. The Institution axis measures AI-specific machinery, not general office capacity, so a long-established office can score poorly when it has withdrawn its AI guidance and imposes no disclosure duty. Read that axis as what an office has built for AI, nothing wider.
Each row below carries the test we applied and the finding behind the score, with a source. The rationale, not the number, is the point.
Scores verified 16 July 2026, against AIP Index codebook v1.0. A score is only true as of its date.
What we look forA granted patent or a clearly workable examination route, not merely a silent statute.
The findingThe 2024 patent amendment admits AI as a computer-implemented invention with a human inventor, and DJKI applies it, but no grant decision demonstrates the AI-assisted pattern.
What we look forProtection applied where human choices shape the work, and refusal of purely machine output.
The findingDJKI protects AI-assisted works with substantial human contribution and refuses purely AI output, through office practice rather than a settled rule, with the codifying bill still in draft.
What we look forA statutory or judicial mining lane. Silence scores against the miner, since the reserved-rights default governs.
The findingThere is no mining exception, and the draft bill would require permission and compensation for training data. With copyright silent, the reserved-rights default governs.
What we look forA traceable, substantive submission on the record, not bare membership.
The findingDJKI engages WIPO as a member, with no traceable substantive Conversation submission found.
What we look forA standalone AI examination text, not AI handled quietly under general practice.
The findingOfficial statements, training material and a bill consultation, but no formal published AI examination guidance.
What we look forA mandatory declaration inside the filing, not a proposal or an informal request.
The findingThe draft bill would require AI-content labeling and DJKI's practice asks for human-contribution evidence, but neither is yet a binding filing duty.
Method. Six questions, three per axis, scored 0, 0.5 or 1 against AIP Index codebook v1.0. The three questions in each axis are weighted equally by choice, and the two axes are reported as a pair, never blended into one number. Every score cites a statute, case or office page, carries a verified date, and records its direction of travel apart from the level. Positions are editorial judgements read from primary law and office practice, not official scores. WIPO's public catalogue of IP-office AI initiatives does not score jurisdictions on this basis, so the readings here come straight from the source law. Not legal advice.
Indonesia's approach to AI is real but unstructured: a thin layer of binding rules over a much thicker layer of ethics circulars, codes of conduct, and drafts. Reviewing it means piecing the picture together from scattered instruments, most of which can't be enforced.
It does three things. It maps the governance structure as an enforceability spectrum, with every instrument linked to its source, so the whole picture reads at a glance. It runs an observatory on the narrow band that actually binds anyone, starting with Article 47 of Permendag 19/2026, to see whether a binding rule with no sanction behind it changes what platforms do. And it tracks the court cases where AI is part of the facts.
A comparative thread runs alongside: where Indonesia states a labelling duty without a method, jurisdictions such as China have already operationalised one. That comparison is where the next round of reform has to work.
Seen a platform using AI, with or without a disclosure label? Send a link and a screenshot, and it will be reviewed before publishing.
Submit an observationWhen a platform uses AI, the job of governing that use falls on it, and so does the choice of how. That is private ordering: rules written and enforced by companies, not by the state. It can move fast, and from the outside it is hard to see. Most people never learn how a platform writes or applies its own AI rules, and they get no say in them.
This site exists to change that. It records how each platform governs its own AI and checks that practice against Article 47 of Permendag 19/2026, so that decisions made inside a company become visible to the public.