Palo Alto Networks announced the launch of its Prisma AIRS platform on a local Singapore cloud. This initiative aims to provide Singapore-based and regional customers with robust AI security solutions tailored to meet data residency requirements while ensuring high performance and security.
The introduction of Prisma AIRS comes at a crucial time for enterprises in the region, where a significant cybersecurity skills gap of 4.8 million workers exists, exacerbated by the increasing reliance on autonomous AI agents. These agents, designed to enhance operational efficiency, also present new security challenges.
The Prisma AIRS platform is positioned to address these challenges by securing privileged access and APIs vital for these AI agents to operate effectively.
Hoseb Dermanilian, regional vice president for AI Security Go-to-Market at Palo Alto Networks, emphasised the need for advanced protective measures.
Hoseb Dermanilian
“As adversaries pivot toward agent-specific attacks, such as prompt injection and tool misuse, a dedicated 'circuit breaker' layer has become non-negotiable. Prisma AIRS addresses this head-on, providing continuous discovery and real-time runtime protection necessary to safeguard the transition to autonomy without introducing severe organisational vulnerabilities.” Hoseb Dermanilian
The platform seamlessly integrates deep model inspection during development with real-time defence during production runtime, ensuring continuous protection across all interactions involving AI models, data, and users.
This approach not only mitigates risks associated with prompt injection and memory manipulation but also ensures that AI innovations are designed with security as a fundamental principle.
Simon Green
“Deepening our investment in Singapore with local hosting of Prisma AIRS is a strategic move to empower regional organizations to deploy AI bravely and at scale," said Simon Green, President of Asia-Pacific and Japan at Palo Alto Networks. "As enterprises across Asia-Pacific and Japan embrace autonomous AI, security cannot be an afterthought.”
Green highlighted that this development aligns with Singapore’s National AI Strategy 2.0, fostering operational resilience while maintaining stringent security standards.
Prisma AIRS encompasses comprehensive end-to-end AI security, featuring three enhanced security modules designed to bolster AI operations:
AI Model Security: This module enables organisations to safely adopt third-party AI models by scanning for vulnerabilities, thus protecting against risks like model tampering and malicious scripts.
AI Red Teaming: Automated penetration testing of AI applications and models is conducted to identify potential vulnerabilities before threats emerge, simulating real attacker behaviour.
AI Runtime Security: This feature safeguards AI applications and models against runtime threats, including prompt injection, malicious code, and sensitive data leaks.
The local availability of Prisma AIRS equips Singaporean organisations with critical AI security capabilities, reinforcing their trust in deploying AI technologies effectively while adhering to data sovereignty obligations.
