Analyst(s): Krista Case
Publication Date: January 15, 2025
Cisco embeds its new AI Defense capability as a part of Cisco Secure Access. The new solution uses AI to support safe and secure AI, specifically targeting security requirements pertaining to access to, and development of, AI applications.
What is Covered in this Article:
- Cisco introduces Cisco AI Defense to address emerging requirements for securing access to AI applications, and securing the AI applications themselves.
- Implications for SecOps and agile application development teams.
- Implications for Red Teams.
The News: Cisco introduces Cisco AI Defense to address emerging requirements for securing access to AI applications, and securing the AI applications themselves.
Cisco AI Defense: Checking the Reckless Charge Toward AI
Analyst Take: Recklessly Charging Toward AI
The world is charging toward AI at a ferocious pace that potentially risks sacrificing safety and security. On one hand, application developers and AI developers and engineers require more speed and flexibility than ever before. On the other hand, this is in direct conflict with the visibility and control that security and IT teams require.
Compounding this dichotomy, AI applications introduce new risk vectors in the form of the AI models that sit between the data and the application. This is creating opportunities for deception and manipulation. For example, we are seeing model poisoning and prompt injection attacks.
Another challenge is that accountability over AI models, AI applications, and their security is fragmented. This will especially be true in 2025 as the number of developers working on AI applications and models continues to boom, and as agentic AI that is taking autonomous action comes into play.
Introducing Cisco AI Defense
In response, Cisco has launched Cisco AI Defense.
In terms of facilitating secure access to AI applications, the solution provides visibility into third-party AI applications in use across the organization, helping to reduce blind spots not only in terms of the totality of AI applications in use, but also what data they are using. It also provides the ability to enforce security policies across this ecosystem. Adoption and the practitioner experience are both streamlined by the offering’s integration into Cisco Secure Access; there is nothing new or additional to install, and security policies can be defined and enforced from a single point. Notably, Cisco identity intelligence is baked into the platform, helping security teams to understand not only which applications are in use and how they are communicating, but also which users specifically are accessing the application, as well as their roles and permissions inside the organization.
To enable organizations to securely develop AI applications, Cisco AI Defense provides visibility into the security of underlying AI models and the data they use. It verifies the integrity of AI models, recommends security guardrails, and can apply those guardrails automatically across public and private cloud environments. This will go far when it comes to balancing the ability to allow developers to retain speed and flexibility, while empowering security teams with control. For example, developers can commit a change, and that change can then automatically validate that it complies with the security guardrails. This is especially useful as AI models, threat tactics, and compliance requirements are constantly evolving.
One of the innovative use cases Cisco shared in an analyst pre-brief call is AI red teaming. The solution provides a readiness score for moving the AI model into production. It identifies the model’s top threats, and then provides recommended guardrails specific to the guardrail across more than 200 categories and utilizing best practices from OWASP, MITRE, and NIST.
For Cisco, this is the next iteration of its focus on embedded software-defined and distributed security capabilities into the network itself. AI Defense Guardrails will be infused into the network fabric to take it close to users and where applications and models are being built. Given Cisco’s incumbency in network infrastructure, the approach is sound.
What to Watch:
- Customer uptick and emerging use cases, given Cisco AI Defense’s integration as a component of Cisco Secure Access.
- How SecOps teams will capitalize on the solution as a function of their CI/CD pipelines, utilizing it for malleable guardrails that facilitate security while not slowing down developers. As an extension of this, how it may facilitate meaningful engagements between security practitioners and developers around AI application architectures and development workflows, and how security fits in.
- How the solution may change the game for Red Teams – uncovering new threats, uncovering threats more quickly, and accelerating time-to-remediation.
See Cisco’s press release for additional detail.
Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.
Other insights from The Futurum Group:
Cisco Bolsters AI Security with Planned Acquisition of Robust Intelligence
Cisco Hypershield: Autonomous, Application-Centric Security
AI Wars: Defending Cyberspace with Intelligence – Six Five Media: Connected
Author Information
With a focus on data security, protection, and management, Krista has a particular focus on how these strategies play out in multi-cloud environments. She brings approximately 15 years of experience providing research and advisory services and creating thought leadership content. Her vantage point spans technology and vendor portfolio developments; customer buying behavior trends; and vendor ecosystems, go-to-market positioning, and business models. Her work has appeared in major publications including eWeek, TechTarget and The Register.
Prior to joining The Futurum Group, Krista led the data protection practice for Evaluator Group and the data center practice of analyst firm Technology Business Research. She also created articles, product analyses, and blogs on all things storage and data protection and management for analyst firm Storage Switzerland and led market intelligence initiatives for media company TechTarget.