🔍 Read the full analysis: A Guide To The Stakes In Indonesia’s AI Copyright Bill on ThorstenMeyerAI.com
TL;DR
Google, Meta and TikTok objected at a November 11 parliamentary hearing to parts of Indonesia’s proposed copyright-law revision, including payments tied to news links and rules for AI training. The draft also addresses AI-generated works, creator-style imitation and disclosure, but it is not law, its current text was not available, and no vote date has been announced.
Google, Meta and TikTok challenged provisions in Indonesia’s proposed copyright-law revision at a parliamentary hearing in Jakarta on November 11, arguing that parts of the bill could affect news access, platform responsibilities and AI development. The proposal would set rules for payments when platforms display news and for AI’s use of copyrighted works, but it remains under discussion and is not law.
The hearing was held by the House of Representatives’ Legislation Body, known as Baleg, which is deliberating the revision with the government. According to Kompas, lawmakers also heard the Press Council and Satu Lisensi Indonesia. The bill revises Copyright Law No. 28 of 2014; parliament approved it as an initiative on March 12, 2026.
The proposal’s reported provisions span two connected disputes. It would require platforms to compensate news publishers for activities including aggregation, republication or showing link previews, with payments routed through state-supervised collective management organisations. It would also establish rules for AI-assisted and AI-generated works, disclosure of AI use, imitation of a creator’s distinctive style, and the use of copyrighted material to train AI models.
Google objected to payment requirements for links and snippets, a single collective licensing body and mandatory licensing for AI training data. It also wants platforms to retain safe-harbor protection. Meta argued that sharing a link is not the same as republishing a story and said payment duties could reduce traffic and revenue for publishers. TikTok said disclosure requirements should not cover cases where AI only assists with a work, and also called for safe-harbor protections to remain, according to the supplied reporting.
AI copyright bill stakes and open questions
At a November 11 hearing, Google, Meta and TikTok challenged parts of Indonesia’s proposed copyright-law revision. The bill is still under discussion, and its current text was not available.
What the proposal could change
The reported provisions affect news publishers, digital platforms, AI developers and creators.
Payments for news links
Platforms could pay publishers for aggregation, republication or link previews, through state-supervised collective management organisations.
Licensing remains disputed
Google opposes mandatory licenses for AI training data and proposes a text-and-data-mining exception with an opt-out.
Authorship and imitation
A July draft described protections for AI-assisted works with human involvement, disclosure rules and a ban on imitating a creator’s distinctive style.
Where the platforms disagree
Their objections focus on payment duties, AI rules and legal protections.
| Company | Position at the hearing |
|---|---|
| Objected to link payments, a single licensing body and mandatory AI-training licenses; sought safe-harbor protection. | |
| Meta | Argued link sharing is not republishing and payment duties could reduce publisher traffic and revenue. |
| TikTok | Opposed disclosure duties when AI only assists with a work and called for safe-harbor protection. |
How deliberations got here
The proposal revises Copyright Law No. 28 of 2014 and builds on an earlier platform-journalism policy.
What remains unsettled
The full text under discussion was not available, so details from the July draft may not match the current proposal.
- How news payments and collective licensing would work
- Whether AI training would require licenses or allow an exception
- How human involvement and style imitation would be defined
- Whether safe-harbor protections or an opt-out would be included
Who Pays for News and AI Training
The proposed law could change how news publishers, digital platforms and AI developers negotiate over content in Indonesia. A payment obligation covering link previews or aggregation would affect how platforms show news and how publishers receive compensation. The bill’s reported plan to route payments through collective management organisations also raises questions about who would set terms and distribute money.
For AI companies, the dispute is whether training on copyrighted works should require licenses, qualify for a fair-use rule, or be addressed through another exception. Google says mandatory licenses for all training data would be technically unworkable at web scale and proposes a text-and-data-mining exception with an opt-out, such as robots.txt. That is the company’s position, not a settled conclusion about what the bill will require or how any rule would work.
The outcome also matters to creators. The draft as reported in July would protect AI-assisted works only when they meet a human-involvement test, while excluding fully AI-generated works. It would also prohibit AI imitation of a creator’s distinctive style. Those provisions could affect creators’ control over their work and the status of AI-supported output, but the available reporting does not establish how the current draft defines the tests or applies them.
Popular right now

Ninja BL660 Professional Blender with 1100-Watt Motor, 72 oz Pitcher, and Two 16 oz Cups

Dyson V9 Motorbar Cordless Vacuum

TCL 55-Inch R7D Series QLED 4K UHD HDR Roku Smart TV (55R7D, 2026)
As an affiliate, we earn on qualifying purchases.
How the Proposal Reached Hearings
The revision is not Indonesia’s first attempt to address the relationship between platforms and journalism. Presidential Regulation No. 32 of 2024 already requires digital platforms to support quality journalism. At the November hearing, Press Council deputy chair Totok Suryanto said the regulation had not worked as intended and platform compliance was low, according to Aktual.
The Press Council’s case rests on the imbalance it says exists between platforms and news organisations. It told lawmakers that platforms account for 80 percent of digital advertising and that traffic from platforms to news sites has fallen by up to 70 percent, according to the supplied source material. These figures were presented by the Press Council; the source does not provide a measurement period or independent verification, so they should not be treated as independently established measures of the bill’s effects.
Reuters reported the AI provisions in a July draft and quoted Hermansyah Siregar, a law ministry official overseeing intellectual property, as saying generative AI had disrupted the copyright framework. Siregar said that draft was not final. The November hearing shows that the text remains contested: companies raised objections, and reporting indicates the draft has changed since July.
„“The development of generative AI has disrupted the copyright framework.”“
— Hermansyah Siregar, a law ministry official overseeing intellectual property, as quoted by Reuters
Key Terms Still Under Debate
The full text being considered on November 11 was not available in the supplied reporting, and accounts refer to different draft versions. Reuters’ July description therefore cannot be assumed to match the current proposal in every detail. The reported references at the hearing to Articles 29 and 30 should also be treated cautiously because the current text could not be checked.
Important questions remain unanswered. Reporting on the July draft did not specify how much human involvement would qualify an AI-assisted work for copyright protection, or how a ban on imitating a creator’s distinctive style would be defined and enforced. The precise terms for news payments, the role and number of collective management organisations, and the scope of any AI-training exception are also unsettled.
It is also unclear whether lawmakers will accept the platforms’ proposals for safe-harbor protections and a text-and-data-mining opt-out. The source material reports no announced vote date, and it does not establish what sanctions would appear in the final bill. Provisions described in an earlier draft, including possible sanctions against platforms, remain subject to change.
Lawmakers Continue Draft Deliberations
Baleg and the government are continuing to deliberate the bill. The next meaningful developments will be an updated version of the draft, further public or parliamentary hearings, and any decision to schedule a vote. No vote date has been announced in the reporting provided.
Until the current text is published or confirmed, readers should distinguish the July draft’s reported provisions from positions stated at the November hearing. Any final assessment of what the law requires will depend on the text approved by parliament and the rules adopted to implement it.
Key Questions
Has Indonesia passed the AI copyright bill?
No. The proposal is still under parliamentary deliberation with the government. The November 11 hearing was part of that process, and no vote date was reported.
What would the proposal require platforms to pay for?
Reporting on the proposal says it would require compensation for activities such as aggregating, republishing or showing previews of news. The precise terms in the version now under discussion remain unclear because its full text was not available.
What AI rules were described in the draft?
A July draft reported by Reuters would address AI-use disclosure, imitation of a creator’s distinctive style, copyright protection for AI-assisted works and use of copyrighted works for model training. That draft was not final, and the current provisions may differ.
What do the platforms want changed?
Google opposes payments for links and snippets, mandatory licensing for all AI training data and a single collective licensing body. Meta says link sharing is not republication, while TikTok says disclosure should not apply when AI only assists. The companies’ positions do not determine the bill’s final terms.
Source: ThorstenMeyerAI.com