Blog
Indonesia's AI compliance wave: Why identity security is the missing piece
Indonesia is moving through 2026 with a distinctive regulatory environment in Southeast Asia. Instead of one comprehensive AI law, AI-relevant obligations are emerging across several regulators, including OJK for financial services, Bank Indonesia for payments, Komdigi for digital platforms, and BSSN for cybersecurity incidents. For any organization deploying AI agents or automated systems, that means a single AI-driven process, say, an algorithm scoring credit risk, can trigger multiple overlapping compliance regimes.
This fragmentation matters right now because enforcement is no longer theoretical. The compliance transition period under Indonesia's Personal Data Protection Law (PDP Law) ended in October 2024, and its implementing regulation, Government Regulation No. 33 of 2026, was issued in July 2026 with its own transition period. A Presidential Regulation on AI ethics remains pending (in January 2026, Communications Minister Meutya Hafid said President Prabowo would soon sign it), and OJK published AI governance guidance for banks in April 2025. The compliance clock has already started. The identity gaps many organizations haven't addressed yet, specifically around who and what is accessing sensitive data, may soon become very visible to regulators.
What's actually changing in Indonesia's AI regulatory landscape
In our view, Indonesia's approach more closely resembles South Korea's Basic AI Act than the EU's staged rollout. Once the Presidential Regulation is signed, individual ministries are expected to issue their own sector-specific rules, meaning obligations will likely land unevenly across industries rather than all at once.
Here's what's already confirmed or in motion:
- A Draft Presidential Regulation on AI Ethics has circulated in 2026 and, as of October 2026, remained under interagency review. As drafted, it would introduce a national ethics framework covering everyone in the AI value chain, from data providers to developers to the operators actually running the systems, building on the 2023 AI ethics circular issued by Kominfo (now Komdigi).
- A companion regulation covering the National AI Roadmap is also expected in 2026, based on a white paper Komdigi released for public consultation in August 2025.
- Pre-AI-law statutory proxies currently govern algorithmic systems through the "electronic agent" concept in Indonesia's Electronic Information and Transactions Law (UU ITE), a framework never built for training-data provenance or autonomous decision-making.
- For banks specifically, OJK's AI Governance for Indonesian Banks guidance (published 29 April 2025) sits alongside existing rules like POJK 11/POJK.03/2022 on IT implementation, creating a compounding compliance stack. It's framed as guidance rather than a standalone binding POJK, but it has been described as the a minimum reference for banks, so treating it as optional could be risky.
The 72-hour breach clock many organizations aren't ready for
Indonesia's PDP Law requires notifying both the PDP agency and affected individuals within 72 hours of a failure of personal data protection, and Government Regulation No. 33 of 2026 ties the start of that clock to when the failure is established. Unlike GDPR, the PDP Law does not set out a general risk-of-harm threshold for notification. If a failure is confirmed, notification will often be the safer course.
That tight window turns into a real problem the moment AI agents enter the picture. If you can't quickly answer what data an identity touched during an incident, human or machine, it becomes much harder to determine your notification obligation in time. And the penalties aren't abstract: the PDP Law's administrative fines can reach up to 2% of annual revenue, and the law separately imposes criminal penalties for certain offenses, such as unlawfully obtaining, disclosing, or using personal data.
Add in data localization requirements under Government Regulation No. 71 of 2019, which mandates local storage for public-scope electronic system operators, and the picture gets more complicated. Industry observers expect the forthcoming AI regulation to reinforce localization specifically for biometric and financial data used in AI training. Any AI agent trained on Indonesian customer data can become a compliance touchpoint.
Why Visibility Is the Real Bottleneck
Indonesia has had real, high-profile incidents that make this concrete. The 2024 ransomware attack on Indonesia's National Data Center disrupted government services including airport immigration processing, and investigators found that a significant portion of the affected data had never been backed up and access paths weren't traceable.
The pattern across incidents like this is the same: identity questions couldn’t be answered fast enough. Who had access? Was it appropriately scoped? Could the access path be reconstructed after the fact? Many organizations don't have that evidence ready, especially when the "identity" in question is an AI agent, a service account, or a bot rather than a person.
This is the governance gap that compliance frameworks keep circling back to: you cannot govern what you cannot see, and you cannot report on an incident you cannot reconstruct.
Treating AI Agents as First-Class, Governed Identities
Closing that gap starts with a shift in how organizations think about AI agents, service accounts, and bots. They aren't background infrastructure. They're identities, and under Indonesia's emerging framework, they need to be governed with the same rigor as human employees.
SailPoint® Agentic Fabric™ was designed to support this shift. It extends SailPoint's enterprise-grade identity security platform to help organizations gain visibility, governance, and automated response across AI agents and non-human identities, as part of one unified system.
In practice, that means:
- Continuous discovery of AI agents, bots, and machine accounts across supported agentic platforms, browsers, endpoints, and SaaS applications, mapped into a single identity graph that shows ownership and data relationships.
- Support for least-privilege and Zero Standing Privilege models, with automated provisioning and deprovisioning designed so agents get access only when needed, and only for as long as they need it.
- An audit trail treating each discovered AI agent as a governed identity, which can assist with goals towards producing the kind of continuous compliance evidence Indonesia's regulators, from OJK to the eventual PDP agency, may expect.
- Machine-speed threat response, with the ability to automatically revoke entitlements or disable compromised agents when anomalous behavior is detected, reducing reliance on manual review.
- Data access governance that helps map the human-to-agent-to-data path, helping close the kind of visibility gap that can make an incident harder to contain and assess.
Instead of leaving compliance evidence scattered across periodic reviews and disconnected tools, Agentic Fabric is designed to help keep it current, continuous, and ready to produce, whether the audience is an internal risk committee or OJK examiners.
Get Ahead of the Regulation, Not Behind It
Today's education-first enforcement posture in Indonesia may not last as the PDP Law's implementing regulation takes effect. Organizations that wait for the Presidential Regulation to be signed before addressing AI agent governance risk building compliance evidence retroactively, under a 72-hour clock, with regulators watching more closely than they are today.
To see how identity security fits into the year ahead for AI governance, download SailPoint's The year of the AI agent report for a closer look at the risks, regulations, and controls shaping the agentic enterprise.
DISCLAIMER: The information contained herein is for informational purposes only, and nothing conveyed herein is intended to constitute any form of legal advice. SailPoint cannot give such advice and recommends that you contact legal counsel regarding applicable legal issues. The development, release, and timing of any features, functionality, capabilities, or services described for SailPoint's products that are not currently available remain at SailPoint's sole discretion on a when, and if, available basis and may not be delivered at all and must not be relied on in making a purchasing or investing decision.