Blog
AI Governance in the Philippines: Why Identity is the compliance control point regulators aren't waiting for
If you're running AI initiatives in the Philippines and waiting for a dedicated AI law before you formalize governance, you're already behind. The Data Privacy Act of 2012 (DPA), enforced by the National Privacy Commission (NPC), already applies to AI systems that process personal data, including in training, testing, or deployment. There's no legislative gap to rely on here; the NPC closed it.
That's the core answer to the question every compliance and security leader in the market should be asking right now: how do we govern AI under Philippine law today, not someday? The mechanism isn't a future statute. One practical mechanism is identity governance applied to systems, humans or machines, that process personal data. Dozens of AI bills are still working through Congress, but the NPC isn't waiting for them, and neither should you.
The Philippines is regulating AI through data privacy law, not AI law
NPC Advisory No. 2024-04, issued 19 December 2024, is the clearest statement so far: AI systems processing personal data fall squarely within the DPA. Before processing personal data to develop or deploy an AI system, including for training and testing, organizations must determine a lawful basis for that processing under Sections 12 and 13 of the Act.
This matters because it collapses two workstreams that many organizations still treat separately, AI oversight and data privacy compliance. In the Philippines, they're closely linked.
The NPC has kept building on that foundation:
- NPC Notice (11 August 2026), using a real person's face or likeness in AI-generated media counts as processing personal data, extending obligations into marketing and media use cases.
- NPC Advisory 2026-01, scraping publicly available personal data for AI training still has to comply with the DPA. "Public" doesn't mean "unregulated."
- NPC Advisory 2026-02 tightened data breach notification procedures, requiring a full report within five days of discovery.
Meanwhile, Congress is anything but idle. The 20th Congress has roughly 40 House bills and 7 Senate bills touching AI in some form, from the Senate's AI Regulation Act (AIRA) to House proposals for a Bureau of Artificial Intelligence Systems, an AI Development Authority, and a Philippine Council on AI.1 None of them are law yet. Compliance programs that treat any single one as an active requirement are getting ahead of the facts, but the volume of activity is a reasonable signal that Philippine AI law may be coming faster than in many neighboring markets. Waiting for it to land before building governance carries significant risk.
Why identity governance can be the practical enforcement point
Here's the connective tissue regulators and threat actors are both circling: identity. The Philippines' cybersecurity market has flagged Identity and Access Management as a fast-growing sub-segment, driven partly by zero-trust adoption and partly by compliance pressure from sector regulators. That's not a coincidence, identity is how "who processed this data, and under what authority" gets answered.
And the threat data backs up the urgency. Viettel Cyber Security recorded more than 52 million exposed credentials in the Philippines in Q3 2025 alone, part of a 49 percent quarter-on-quarter rise in recorded data breaches.2
When your compliance posture depends on proving who, or what, had valid access at any given moment, identity governance isn't a supporting control. It's a central control.
The AI agent blind spot regulators haven't caught up to yet
The DPA and the NPC's AI guidance were written with data processing in mind, not specifically with autonomous AI agents that independently execute tasks, call APIs, and touch sensitive data without a human clicking "approve" each time. But the underlying obligation doesn't change: if an agent processes personal data, someone has to be accountable for the lawful basis, the access it holds, and what it did with that access.
That's a governance problem some identity programs weren't built for. AI agents and machine identities can vastly outnumber human employees in many enterprises, and they operate at machine speed. Traditional, human-centric access reviews, designed around quarterly or annual cycles, can leave months-long gaps where an agent's access goes unvalidated. In a regulatory environment where the NPC is tightening breach notification timelines, that may be a point of regulatory scrutiny.
What continuous AI governance actually requires
Meeting the DPA's transparency, accountability, and lawful-basis requirements for AI systems means an organization needs to be able to answer:
- Which AI agents and machine identities exist across the enterprise, including shadow AI tools nobody formally provisioned.
- Who owns each agent and what data or systems it can reach.
- Whether that access reflects least privilege, or whether it's standing access nobody's revisited since deployment.
- What happened when an agent behaved unexpectedly, and how fast the organization responded.
- Whether there's an audit trail that holds up when the NPC comes calling.
This is where SailPoint® Agentic Fabric™ fits, as an identity layer designed to help make those five questions answerable on demand. It extends SailPoint's enterprise-grade identity security platform, whose customers represent 53% of the Fortune 500, to treat AI agents and machine identities as first-class governed identities rather than an invisible risk.
Agentic Fabric is designed to continuously discover and register agents across agentic platforms, browsers, endpoints, and SaaS applications, mapping each one to its human owner, its entitlements, and the sensitive data it can touch. It helps enforce least privilege and Zero Standing Privilege through automated, just-in-time provisioning instead of standing access that lingers unreviewed. And because compromised agents don't wait for a quarterly review to cause damage, Agentic Fabric monitors behavior, scores risk dynamically, and can trigger automated remediation, revoking entitlements or disabling an agent when anomalous activity is detected.
Just as importantly for a regulatory environment defined by frequent NPC advisories rather than a single static law, Agentic Fabric is designed to maintain a continuous, policy-driven audit trail. That helps keep compliance evidence from going stale between review cycles, because governance isn't a periodic event, it's constant.
Build governance that doesn't depend on a law that hasn't passed yet
The Philippines' regulatory trajectory is clear even without an enacted AI statute: obligations are expanding, enforcement infrastructure is scaling, and identity is the thread connecting data privacy compliance, sector-specific guidance, and breach prevention. Organizations that wait for AIRA or its House counterparts to become law before formalizing AI governance are choosing to govern reactively in a market where regulators are already acting proactively.
Want to see where AI agent adoption and governance are heading before the rest of the market catches up? Download our The year of the AI agent report for a data-driven look at how enterprises are securing the agentic workforce today.
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.
Sources
- Rappler, “How AI governance is taking shape in the Philippines,” 6 June 2026 (about 40 House bills and 7 Senate bills in the 20th Congress), via DPEX Network; bill details: Foundation for Media Alternatives.
- Viettel Cyber Security, Q3 2025 Cyber Threat Landscape Report, as reported by Context.ph (24 October 2025).