TLDR: Singaporean organizations demonstrate leading capabilities in AI security, with 52% reporting full preparedness, exceeding the global average. However, a significant 62% of these firms are frequently battling ‘shadow AI’ threats, highlighting a critical gap between perceived readiness and actual vulnerability.
A recent Delinea research report, ‘AI in Identity Security Demands a New Playbook,’ reveals a compelling paradox within Singapore’s cybersecurity landscape. While Singaporean organizations are at the forefront of AI security preparedness, with 52% indicating they are fully equipped to secure AI technologies – a figure notably higher than the global average of 44% – they simultaneously face the highest global incidence of ‘shadow AI’ threats. A staggering 62% of Singaporean businesses report encountering challenges from unsanctioned AI use at least once a month, underscoring a significant disconnect between confidence and operational reality.
The report, published on September 4, 2025, highlights that despite 97% of Singaporean organizations believing their machine identity security can effectively manage AI-driven threats, only 70% possess comprehensive visibility into these machine identities. Furthermore, a mere 56% have implemented specific governance policies for AI identities. This oversight is particularly concerning given the widespread adoption of agentic and generative AI, with 91% of Singaporean businesses utilizing these technologies daily. The limited visibility and inadequate governance leave these AI agents susceptible to unchecked autonomous decisions and potential external compromises.
Art Gilliland, CEO of Delinea, emphasized the urgent need for ‘agentic security’ to match the demands of ‘agentic AI.’ He urged organizations to re-evaluate their identity approaches, advocating for adaptive, risk-aware systems capable of verifying and securing every action, regardless of whether it originates from a human or a machine. Gilliland stressed that AI agents necessitate more granular and dynamic identity access controls than traditional role-based methods. He also underscored the importance of establishing a comprehensive AI governance model to ensure secure and intended AI usage.
Also Read:
- Securing Non-Human Identities: The Challenge of AI Governance in the Enterprise
- ASEAN’s Evolving AI Governance: Navigating a Diverse Regulatory Landscape
Key findings from the report pinpoint several critical concerns for respondents. AI-generated phishing and deep fakes are the top worry for 54% of firms, closely followed by agentic AI systems operating with unchecked access (52%), and AI-driven credential theft (47%). The governance deficit is further evidenced by the fact that only 63% of organizations have an acceptable use policy for AI tools, and just 64% have access controls in place for AI agents. These findings collectively highlight the imperative for organizations to prioritize identity governance and implement adaptive security controls to effectively mitigate the inherent risks associated with advanced AI technologies.
