Singapore organisations surveyed reported average losses of US$3.49 million from insider security incidents over the past 12 months—the highest among 14 countries included in Trellix’s Mind of the CISO: The New Insider research. The findings point to AI and autonomous systems as an emerging insider-risk vector for enterprises across Asia.
Among Singapore CISOs surveyed, 84% are concerned about sensitive-data leakage caused by AI-driven or autonomous systems, while 60% say they lack sufficient visibility into how employees use AI tools within their networks.
Globally, Trellix found that 95% of CISOs and senior security leaders experienced an insider incident linked to an AI or autonomous system in the past 12 months. AI-related actions accounted for 28% of all insider incidents on average, while the global average cost was US$2.4 million.
The research highlights a widening gap between the speed at which AI operates inside enterprises and the speed at which security teams can detect and respond to the risks it introduces. Trellix says adversaries are commercialising AI-enhanced attack tools and exploiting vulnerabilities faster than human-speed defences can respond.
“Across Asia Pacific, adversaries are already weaponising AI to outpace human defenders, and security teams need to move quickly to match the speed and scale of attacks,” said Neville Vincent, senior vice president and general nanager, APJC, at Trellix. “The answer cannot be AI that demands a trade-off between capability and control.”
To address that gap, Trellix has launched Trellix Wise for Private Cloud, an on-premises, federated AI core for agentic security operations centres. It can be deployed in on-premises, air-gapped, hybrid or private-cloud environments, with data remaining inside the customer’s security boundary rather than being sent to a public large language model or external cloud.
The platform connects telemetry from AI agents and traditional security sources across existing infrastructure. Trellix says it is designed to investigate all received alerts, determine whether alert clusters represent genuine threats, and provide auditable explanations, confidence levels and recommended response actions.
For Singapore organisations, 40% plan to prioritise detection and response investment for AI-related identity risks over the next 12 to 24 months. This reflects the growing importance of governing AI agents, machine identities and privileged access alongside conventional user accounts.
