As AI adoption accelerates, identity is becoming Singapore’s most contested cybersecurity battleground. 

Singapore stands on the cusp of a new digital inflection point. With AI scaling across industries, identity emerging as the primary security perimeter, and enterprises modernising at record speed, the country is more connected and more vulnerable than ever before. Identity is no longer just a way to authenticate users; it has become the most actively targeted attack surface and the point that determines the blast radius of any breach. 

This risk is magnified by the pace and shape of digital adoption in Singapore. IDC estimates that 66 per cent of APAC enterprises will embed AI-driven decisioning into core business processes over the next three to five years, while Gartner projects that non-human identities such as workloads, bots, and AI agents will outnumber human identities by at least 45 to 1 in digitally mature markets like Singapore.

Here’s why Southeast Asia’s infrastructure startups are quietly becoming the region’s most valuable companies

Against this backdrop, six identity-driven challenges are particularly critical for Singapore’s BFSI, healthcare, manufacturing, government, and GCC ecosystems as organisations get ready for 2026. These dangers are already present and will have a big impact on how businesses in Singapore develop cyber resilience going forward. 

Agent Identity Hijacking 

The rapid adoption of GenAI has created a new type of identity inside enterprises: autonomous agents that draft content, interact with systems and, increasingly, make decisions. These agents operate with entitlements, just like human users. When they are compromised, attackers can use them with focused accuracy and speed without being discovered immediately. A hijacked agent can move laterally across systems, execute privileged actions, or leak sensitive data without the typical fingerprints left by human attackers.  

The threat is especially salient in Singapore and across the Asia Pacific, where enterprises are advancing quickly from experimenting with to the actual use of autonomous AI. In fact, IDC found that around 70 per cent of APAC organisations expect agentic AI to disrupt business models within the next 18 months, and 34 per cent now prioritise AI governance as they scale GenAI use, underscoring just how rapidly these systems are being embedded into mission-critical processes. 

With Singapore enterprises embracing AI copilots across IT services, banking, healthcare and customer operations, agent identity governance will become essential for managing emerging risk. 

Prompt injection and intent manipulation 

Prompt injection has become the next evolution of phishing, except the target is now the AI model itself. By embedding malicious instructions in prompts, forms or unstructured content, attackers can coerce an enterprise AI system into behaving incorrectly or dangerously. This can lead to exposure of sensitive data, manipulation of workflows or bypassing of controls. In Singapore, where AI-driven chat, decisioning and support systems are being deployed at scale, even subtle manipulation of model intent can trigger large-scale impact, IDC’s Asia Pacific Security Study found that 76.5 per cent of enterprises in the region say that they are not confident in their organisation’s ability to detect and respond to AI-powered attacks, reflective of a growing trust gap as AI scales into critical systems. Securing intent even more than securing user inputs will be critical as AI becomes the interface for enterprise processes. 

Model poisoning 

Model poisoning attacks aim to corrupt the training data or fine-tuning pipelines that power enterprise AI systems. For identity and security systems, this is particularly damaging. A poisoned model may approve risky access, fail to flag anomalies or misjudge behavioural risk. Across the Asia Pacific, AI and machine learning have become central to fraud detection and related identity workflows in BFSI and healthcare, with 90 per cent of financial institutions now using AI to fight fraud and financial crime. Protecting against model poisoning in Singapore will require visibility into training data sources, stringent controls on model updates and continuous validation of AI behaviour against expected patterns. 

Shadow entitlements 

Untracked or hidden permits, sometimes known as ‘shadow entitlements,’ are a silent but growing concern. With the rapid use of SaaS platforms, cloud services, DevOps pipelines, and automation technologies, firms in Singapore are getting access to privileges not seen in traditional workflows. These privileges are frequently generated automatically, abandoned following system integrations, or given momentarily and never revoked. Because they are invisible to traditional audits, reducing identity risk will depend on controlling entitlement sprawl through automated cleanup and ongoing discovery as digital ecosystems grow. 

Supply-chain identity compromises  

The next generation of supply-chain attacks will see an increased focus on identification rather than software. Attackers have begun to increasingly focus on contractors, vendors, cloud partners, and third-party users – identities that frequently have permanent or privileged access to enterprise systems. This risk is heightened in Singapore and across the Asia Pacific, where enterprises operate within deeply interconnected, outsourced ecosystems. As reliance on partners grows, attackers can exploit third-party identities to get direct access to crucial areas. Applying zero-trust principles uniformly across every external identity will be critical for protecting the integrity of enterprise operations. 

One thing unites all five of these new risks: just as identity is changing, so are the threats that surround it. Humans, machines, AI agents, contractors, and autonomous systems are all part of this constantly changing ecology. Implementing continuous risk-based authentication, removing standing privileges, strictly regulating both human and non-human identities, and improving detection utilising real-time, AI-driven insights are all essential to safeguarding Singapore’s digital future. 

Organisations that fortify their identity foundations today will be in the best position to protect trust, maintain operational continuity, and deal with the challenges of an increasingly complex digital ecosystem as Singapore gets closer to 2026. 

The article titled “The identity threats Singapore must prepare for in 2026” was authored by Marco Zhang, Director, Solutions Engineering, Saviynt

About the author

Marco Zhang is Solutions Engineering Director for Asia Pacific and Japan at Saviynt, where he advises enterprises on strengthening identity security and managing identity risk across hybrid and cloud environments. With more than 20 years of experience in cybersecurity and enterprise infrastructure, he specialises in identity governance, privileged access management, and Identity Security Posture Management (ISPM).

Before joining Saviynt, Marco held senior roles at Oracle, Sun Microsystems, Dell Software, Dimension Data, and identity security firms, including One Identity, BeyondTrust, and Centrify, supporting large-scale security and identity programs across Asia Pacific. At Saviynt, he leads the regional sales engineering team and works closely with organisations to modernise identity programs and strengthen their security posture.

Marco holds master’s degrees in Mechanical Engineering and Communication Software & Networks, and maintains CISSP and PMP certifications.