F5 has announced the availability of F5 AI Guardrails and F5 AI Red Teamtwo solutions tasked with protecting mission-critical enterprise AI systems. With this step, F5 offers a comprehensive, end-to-end lifecycle approach to runtime AI security, including more ability to connect and protect AI agents with custom, out-of-the-box guardrails.
These security solutions adapt to the needs of enterprises, offering flexible deployment, model-agnostic protection, and the ability to adapt AI security policies in real time. To do this, they take advantage of F5’s experience in the application layer, where AI interactions occur.
F5 AI Guardrails offers a model-independent runtime security layer. It is designed to protect AI models, applications, and agents across any cloud environment and deployment with consistent policy enforcement. This solution offers protection against adversary threats such as prompt injection and jailbreak attacks.
Additionally, it prevents the leak of sensitive data and complies with corporate and regulatory obligations, including the GDPR and the EU AI Law. AI Guardrails also offers comprehensive observability and auditability of AI inputs and outputs so teams can see, in addition to what the model did, why it did it. This is fundamental for governance and regulatory compliance in regulated sectors.
As for F5 AI Red Team, it offers scalable and automated adversarial tests that simulate common and unknown threat vectors that appear in the industry’s AI vulnerability database (this database adds more than 10,000 new attack techniques per month). AI Red Team reveals where models can generate dangerous or unpredictable results.
Their insights are fed directly into AI Guardrails policies, with the goal that defenses evolve as threats and models themselves change. In combination, AI Guardrails and Red Team establish a Continuous AI Safety Feedback Loop. Both are now available as part of the F5 Application Delivery and Security Platform.
