Presidential Regulation on the Framework for Artificial Intelligence Development

· About Presidential Regulation on the Framework for Artificial Intelligence Development

Key Takeaways

  • Indonesia’s national AI framework is not yet in force as a presidential regulation as of 4 October 2026; the government had completed drafting of the AI roadmap and AI ethics regulation in July 2026, but public reporting still describes them as pending promulgation.[11][14]
  • The proposed regime is intended to apply across government, industry, academia, the public, and media, with a national governance model led by the Ministry of Communication and Digital Affairs and coordination across ministries and agencies.[10][14]
  • The draft policy package is expected to cover AI roadmap governance, ethics and safety, and sectoral follow-on rules, including transparency measures such as AI-content labeling or watermarking.[2][10]
  • The current materials do not show enacted penalties, fines, or binding phase-in deadlines for the draft presidential regulation itself; those sanctions remain uncertain until the regulation is formally issued.[11][14]
  • Indonesia is positioning the framework as an executive, risk-based AI regime tied to national interests, cultural norms, and responsible innovation, rather than as a single comprehensive AI statute.[10][12]
  • Organizations operating in Indonesia should treat the draft as a near-term compliance signal and align AI governance now to reduce implementation risk once the presidential regulation is signed.[7][14]

What It Is

Indonesia’s AI framework is a proposed presidential regulation package rather than an enacted statute, centered on two linked instruments: a National AI Roadmap and an AI Ethics / safety framework.[10][14] Public reporting from mid-2025 through September 2026 says the Ministry of Communication and Digital Affairs has been drafting the package, with the government aiming to move it through harmonization and then to presidential signature.[4][11]

The available sources indicate the policy is national in scope and meant to shape AI development and deployment across sectors, including public administration, law enforcement, finance, education, health, and the creative economy.[9][10] The Ministry of Communication and Digital Affairs appears to be the main policy lead, while inter-agency coordination is expected through a national task-force or similar mechanism.[10][14]

The exact adoption date and formal in-force date are not yet available because the instrument remains in draft form in the latest reporting.[11][14] The most specific 2025–2026 milestone visible in public reporting is that the government targeted harmonization by late September 2025 and later described the draft package as a 2026 priority.[4][7] As of July–October 2026, credible reporting still describes the AI roadmap and ethics regulation as being finalized or awaiting promulgation.[2][11][14]

Who Must Comply

The draft regime is described as a national framework intended to guide government, industry, academia, the public, and the media in developing and using AI in Indonesia.[10] That broad framing suggests it is not limited to licensed AI vendors or public bodies alone.[10][14]

Because the draft is still pending, no final applicability thresholds have been publicly confirmed. The reporting does not yet identify a formal revenue threshold, employee-count threshold, or system-volume threshold comparable to the EU AI Act.[11][14]

Extraterritorial reach is also not settled in the sources reviewed. However, a framework built around national governance, content safety, and sectoral implementation would likely affect foreign providers whose systems are offered in Indonesia or integrated into Indonesian-facing services, especially if later rules impose labeling, transparency, registration, or sector-specific approval obligations.[10][12]

No specific exemption regime has been confirmed in the publicly reported material. If exemptions are included in the final text, they may track low-risk internal use, research, or narrow public-interest carve-outs, but that remains speculative until the presidential regulation is issued.[11][14]

Core Requirements

  1. National roadmap alignment — The framework is expected to require ministries and agencies to align AI projects with the national roadmap and any sectoral follow-on rules, making AI governance a state-coordinated exercise rather than an ad hoc procurement issue.[9][10]
  1. Ethics and safety governance — The draft ethics instrument is intended to establish a binding basis for safe, responsible AI development, deployment, and use, with attention to transparency, accountability, and risk control.[10][14]
  1. Transparency measures — Public reporting indicates the first implementing measures may include AI-generated content labeling or watermarking, especially for digital-content contexts.[2]
  1. Inter-agency oversight — The emerging model includes a coordinating role for the Ministry of Communication and Digital Affairs and cross-ministry or cross-agency governance structures to supervise AI adoption across sectors.[10][14]
  1. Sectoral implementation — The national framework is expected to push ministries and agencies to issue more detailed sector rules after the presidential regulation is signed, rather than resolving every use case in one instrument.[2][9]
  1. Local values and national interest — The stated policy direction emphasizes ethical standards, national interests, and conformity with local values and cultural norms, which suggests content, labeling, and deployment decisions may be reviewed through a public-interest lens.[10][14]

Deadlines and Penalties

| milestone | date | what applies | |---|---:|---| | Draft targeted for harmonization | 25 September 2025 | Government aimed to begin harmonization and legal review of the AI draft.[4] | | AI regulation made a 2026 priority | 26 January 2026 | Ministry leadership said the AI roadmap and ethics regulations were a priority for presidential signature in 2026.[7] | | Draft completed and awaiting promulgation | 27 July 2026 | Reporting said the government had completed drafting the AI roadmap and AI ethics presidential regulation.[11] | | Status at latest reporting | 4 October 2026 | No confirmed in-force presidential regulation text was identified in the sources reviewed.[11][14] |

Because the instrument remains draft, maximum fines and other sanctions are not yet confirmed in the publicly available sources used here.[11][14] At present, the safest statement is that sanctions are unknown until the regulation is promulgated.

How to Comply

  1. Map AI use cases now — Inventory models, vendors, and internal use cases, then classify them by business function, user impact, and data sensitivity to see what would likely fall under national AI governance once the regulation is enacted.
  1. Build an AI governance register — Use ISO 42001 as the management-system backbone for policy, roles, incident escalation, documentation, and continuous improvement, because the Indonesian draft is aimed at repeatable governance rather than one-off approvals.
  1. Align controls to ISO 27001 and NIST CSF 2.0 — Map AI data, access, supplier, logging, and incident-response controls to ISO 27001 and NIST CSF 2.0 so the organization can demonstrate a mature security baseline if the final regulation requires safety or traceability.
  1. Prepare content provenance controls — If the final text includes watermarking or labeling, establish technical and editorial workflows for AI-generated content identification, approval, and retention.
  1. Run risk and impact reviews — Add pre-deployment impact assessment for higher-risk AI uses, including bias, safety, explainability, consumer harm, and public-sector impacts, with sign-off from legal, security, privacy, and product owners.
  1. Review data governance and localization dependencies — Verify lawful bases, cross-border transfer positions, retention schedules, and sensitive-data handling because AI governance in Indonesia is likely to overlap with data protection and sectoral compliance.
  1. Prepare for sectoral rules — Track ministry-level guidance for regulated sectors such as finance, health, education, and law enforcement, since the national framework is expected to trigger more detailed implementing rules.

Related Regulations

  • Indonesia Personal Data Protection Law — AI systems that ingest or generate personal data will still need a lawful processing basis, purpose limitation, and security controls, so the AI framework is likely to sit on top of PDP compliance rather than replace it.
  • Electronic Information and Transactions framework — Content, platform, and electronic-system obligations may overlap with AI-generated content rules, especially where the final regulation addresses labeling, transparency, or harmful outputs.
  • Sectoral financial-services rules — AI used in lending, fraud detection, or automated decision-making will likely be constrained by financial regulator expectations on governance, model risk, and consumer protection.
  • Cybersecurity and electronic-system rules — AI governance will overlap with system integrity, logging, incident response, and resilience obligations already expected of operators of electronic systems.
  • Creative economy and content rules — If the final presidential regulation addresses AI-generated media, it may interact with copyright, content moderation, and creative-industry policy.

FAQ

Does the AI presidential regulation already apply in Indonesia?

Not yet, based on the latest sources reviewed. Reporting in July through October 2026 still describes the AI roadmap and ethics regulation as draft or in finalization, not as a promulgated and in-force presidential regulation.[11][14]

Does it apply to companies outside Indonesia?

The final extraterritorial scope is not confirmed because the regulation is still in draft. If the text follows the government’s stated aim of governing AI use in Indonesia, foreign providers offering AI-enabled services into the Indonesian market are likely to be affected once implementation rules are issued.[10][14]

Will AI-generated content need a label or watermark?

That appears to be one of the first likely implementing measures. Public reporting in 2026 says the ministry was preparing content-labeling or watermarking rules as a transparency measure for AI-generated material.[2]

What penalties apply for non-compliance?

No verified penalty schedule is available yet for the draft presidential regulation. Until the regulation is formally issued, any fines, suspension powers, or administrative sanctions remain uncertain.[11][14]

Is this the same as Indonesia’s data protection law?

No. The AI framework is a governance and ethics instrument for AI development and use, while Indonesia’s data protection law governs personal-data processing; organizations will need to comply with both where AI systems use personal data.

What should multinationals do now?

Multinationals should map Indonesian AI use cases, document governance, and prepare for labeling, risk assessment, and sector-specific controls. ISO 42001, ISO 27001, and NIST CSF 2.0 provide practical structures for readiness while the final text is still pending.

Sources

Put it into practice

More compliance guides