Cisco unveils AI Defense to secure enterprise AI transformation

Source: Cisco. An overview of Cisco AI Defense capabilities.

Cisco, the security and networking provider, has announced Cisco AI Defense to enable and safeguard AI transformation within enterprises. As AI technology advances, new safety concerns and security threats are emerging at an unprecedented speed, too fast for existing security solutions to manage, the company said.

Cisco AI Defense is purpose-built for enterprises to develop, deploy and secure AI applications with confidence. "Business and technology leaders can’t afford to sacrifice safety for speed when embracing AI,” said Jeetu Patel, Executive VP and Chief Product Officer, Cisco.

"In a dynamic landscape where competition is fierce, speed decides the winners. Fused into the fabric of the network, Cisco AI Defense combines the unique ability to detect and protect against threats when developing and accessing AI applications without tradeoffs.”

Raymond Janse van Rensburg, VP Networking Sales and Solutions Engineering Asia-Pacific, Japan, and Greater China, Cisco, noted that architectures are no longer as simple as applications linking to data and hardware infrastructure when it comes to AI. In a multi-model, multicloud world, output can be hard to predict, and there are a new set of risk vectors, he said. "When these models break, bad things will happen," he said. 

According to Cisco's 2024 AI Readiness Index, only 28% of those surveyed in Singapore feel fully equipped to detect and prevent unauthorised tampering with AI. The security challenges are also new and complex, Cisco pointed out, with AI applications being multi-model and multicloud. Vulnerabilities can occur at model or app level, while responsibility lies with different owners, including developers, end users and vendors.

As enterprises look beyond public data and begin training models on proprietary data, the risks only grow, Cisco said. To unlock AI innovation and adoption, enterprises need a common layer of safety and security that protects every user and every application.  We have to fundamentally reimagine safety and security for the era of AI," van Rensburg said, suggesting that enterprises look for a solution that is part of a security platform instead of a point solution.

AI Defense enables enterprises’ AI transformations by addressing two urgent risks:

Developing and deploying secure AI applications

As AI becomes ubiquitous, enterprises will use and develop hundreds if not thousands of AI applications. Developers need one set of AI security and safety guardrails that work for every application. AI Defense helps developers move fast and unlock greater value by protecting AI systems from attacks and safeguarding model behaviour, across platforms. The capabilities of AI Defense include:

• Discovering AI: Security teams need to understand who is building applications and the training sources they use. AI Defense detects shadow and sanctioned AI applications across public and private clouds.

• Model validation: Model tuning can lead to toxic and unexpected outcomes. Automated testing checks AI models for hundreds of potential safety and security issues. This AI-driven algorithmic red team* identifies potential vulnerabilities and recommends guardrails in AI Defense for security teams to use.

• Runtime security: Continuous validation safeguards against potential safety and security threats such as prompt injection, denial of service and sensitive data leakage on an ongoing basis.

Securing access to AI applications

As end users rush to adopt AI applications like summarisation tools to improve their productivity, security teams need to prevent data leakage and the poisoning of proprietary data. AI Defense enables security teams with:

• Visibility: Provides a comprehensive view of shadow and sanctioned AI-enabled apps used by employees.

• Access control: Implements policies that restrict employee access to unsanctioned AI tools.

• Data and threat protection: Continuously safeguards against threats and confidential data loss while ensuring compliance.

Unlike safety guardrails built into individual AI models, Cisco delivers consistent controls for a multi-model world. AI Defense is self-optimising, leveraging Cisco’s proprietary machine learning models to detect AI safety and security concerns based on threat intelligence data from Cisco Talos. Splunk customers that are using AI Defense will receive enriched alerts with additional context from across the entire ecosystem.

AI Defense integrates with existing data flows for visibility and control and is built into the Security Cloud, Cisco's unified, AI-driven, cross-domain security platform. It leverages Cisco’s extensive mesh of enforcement points to perform AI security at the network level. Accuracy and trustworthiness are essential for protecting enterprise AI applications, and Cisco has been actively involved in developing AI security industry standards, including those from MITRE, OWASP, and NIST.

"The adoption of AI exposes companies to new risks that traditional cybersecurity solutions don't address," said Kent Noyes, Global Head of AI & Cyber Innovation at World Wide Technology, a technology company with a strong Asia-Pacific presence.

"Cisco AI Defense represents a significant leap forward in AI security, providing full visibility of an enterprise's AI assets and protection against evolving threats."

AI Defense is the latest in a series of AI-driven security introductions from Cisco, including Cisco Hypershield. Cisco AI Defense will be available in March, and requires Cisco Secure Access. Early access is available. Early adopters include Deloitte, IBM, Hitachi, Honda, and Sompo.

*A red team is a group of security professionals who attack cyberdefences to evaluate them.