AI has dramatically reshaped the cloud network security landscape. As organizations have implemented GenAI and agentic AI into their operations, two main security issues have come into focus:
Security for AI – Securing AI agents and LLMs from being hacked or misused
AI for security – Using AI to make networks more secure through automation, pattern recognition, threat detection, and analysis
Through our Aviatrix Cloud Native Security Fabric (CNSF) platform, we secure public cloud networks and AI workloads. To reflect our commitment to securing customer networks no matter what AI innovations come next, we're changing our website domain from "aviatrix.com" to "aviatrix.ai." Read on to learn more about how AI introduces new possibilities and risks and how Aviatrix CNSF helps organizations adapt securely.
AI: New Innovations and New Threats
GenAI and agentic AI unlock enormous potential for efficiency and innovations through task automation, knowledge-sharing, data analysis, and content creation. However, AI comes with unprecedented risks:
Threat actors can hijack AI agents that have access to sensitive data
LLM training data can be poisoned to manipulate outputs
Multiple AI agents can collude to evade system security and steal data
Shadow AI can lead to employees entering sensitive data into insecure LLMs
Many organizations are not prepared with robust AI security frameworks, though AI use is increasing rapidly. The Aviatrix State of Cloud Network Security: 2025 report found that organizations are using AI for security: 95% of surveyed organizations leverage artificial intelligence and/or machine learning for threat detection. However, IBM’s Cost of a Data Breach 2025 report found that actually securing GenAI or agentic AI use is not as prevalent: 63% of organizations included in that survey lacked AI governance policies to manage AI or prevent the proliferation of shadow AI. Darktrace found that many threat actors are using AI for email phishing and social engineering attacks.
Aviatrix CNSF answers the challenges of AI security by securing data, systems, and users, leaving organizations free to explore AI innovation safely.
Aviatrix Cloud Native Security Fabric (CNSF): Security for AI, and AI for Security
Cloud Native Security Fabric (CNSF) is a new paradigm for cloud network security that embeds security principles into the fabric of the cloud itself. Aviatrix CNSF provides a holistic, pervasive security solution that protects all cloud accounts, users, and traffic types with consistent and enforceable policies and network-wide visibility.
Here’s how CNSF answers the challenge of AI security:
Security for AI
Aviatrix CNSF enforces zero trust principles for AI workloads, guarding networks from AI vulnerabilities like:
Goal hijacking – Manipulating an AI agent’s goals to control its behavior
Memory poisoning – Injecting false or malicious data into an AI agent’s long-term memory to alter future decisions
Supply chain vulnerabilities – Compromises in model dependencies, libraries, or APIs that can threaten the integrity of LLM data
AI may be relatively new, but core cloud network security tenets remain the same – and Aviatrix leverages those basic principles to protect networks from AI misuse and mistakes.
Ingress and egress filtering
Most organizations focus on ingress security, or preventing malicious actors from getting into their networks. However, cloud networks have a wide attack surface that is much easier to infiltrate. Egress security means filtering outgoing traffic from your network to prevent it from going to suspicious domains or exfiltrating data.
Whether a human user or an AI agent is accessing data for exfiltration, they will need to collect it and remove it from the network. Aviatrix CNSF uses egress filtering to inspect outgoing network traffic and prevent data theft.
Identity-based workloads
AI agents are gaining the power to act like human users, without the limits of identity and access management (IAM) controls or role-based access control (RBAC) boundaries. They can generate access tokens, run scripts, and make API calls on their own. Aviatrix CNSF resolves the security issue of identity by basing it on workload rather than ephemeral values like IP addresses.
Network segmentation
A flat network allows anyone, human user or AI agent, to roam freely once inside. Aviatrix CNSF provides network segmentation that keeps systems separate, meaning that any AI agent or other user can only access one piece of the system. This segmentation limits the blast radius of any incident.
Security policy enforcement
Autonomous AI agents can evade security policies through unsecured integration with systems that have sensitive data. Aviatrix CNSF uses empowers you to set and enforce consistent, identity-aware security policies that align with zero trust principles and span every cloud, location, and environment, closing those loopholes.
Visibility and monitoring
Because AI agents operate across microservices using HTTPS, they avoid firewall traffic inspection and DPL (data prevention loss) tools. Aviatrix CNSF offers inline visibility into all workloads and traffic types, including east-west traffic within your network to prevent lateral movement. It can also inspect command-and-control traffic for unusual task execution, revealing when an AI agent’s goals have been hijacked.
Infrastructure as code (IaC)
Aviatrix integrates with Terraform to empower cloud architects, DevOps, and DevSecOps to design security-first, repeatable network architectures. This integration secures CI/CD pipelines and removes the blind spots and design flaws that poisoned AI agents could exploit.
AI for Security
As well as securing AI workloads, AI uses AI agents and LLMs to enhance cloud network security. Here are a few of the most recent AI advancements:
Aviatrix Secure Network Supervisor with Microsoft Copilot for Security – This feature integrates with Microsoft Copilot for Security’s AI capabilities to troubleshoot complex VPN issues by analyzing root causes and guiding remediation.
Integration with Wiz – Aviatrix partners with Wiz to turn network visibility into instant security policy enforcement. Wiz spots suspicious activity and “toxic combinations” like high-value, vulnerable workloads; Aviatrix takes those insights and uses AI automation to enforce security policies and prevent threat actors from doing any damage.
UX with Explainable AI (XAI) - Aviatrix CNSF uses a UX strategy that gives the human user AI-driven insights about network design, performance, and security policies so they can make informed decisions.
“Aviatrix.ai” signals our commitment to delivering runtime security and enforcement through CNSF, closing blind spots and simplifying complex network management so our customers can scale and grow without compromise. Schedule a demo today to learn more about how CNSF can protect your network in the age of AI.
Discover more about how Aviatrix provides security for AI and AI for security:
Learn about six AI accidents that could cost you everything – and how to prevent them.
Explore how Aviatrix CNSF enables AcrossAI to deliver secure enterprise AI agents.
