Cisco this week unveiled AI Defense, a new solution designed to help enterprises secure the development and use of AI applications. 

Cisco AI Defense focuses on two main areas: accessing AI applications, and building and running AI applications.

The first is related to the use of third-party AI apps, which can boost productivity, but they can also introduce risks, such as potential data leakage and malicious downloads. 

Cisco AI Defense aims to address this by providing full visibility into AI app usage, by offering access control capabilities to restrict access to unsanctioned AI tools, and by providing protection against threats and confidential data loss. 

As for building and running AI applications, the new solution aims to help enterprises by giving them the tools to discover shadow and sanctioned AI applications, by providing automated testing to validate AI models and identify vulnerabilities, and by providing runtime protection against threats such as prompt injection, DoS attacks, and sensitive data leakage. 

“At a high level, the risk of enterprise AI is two-fold. There are practical concerns about sensitive data exposure from employees sharing intellectual property, PII, and other confidential information with unsanctioned third-party AI tools,” explained DJ Sampath, VP of product for Cisco’s AI Software and Platform. “At the same time, businesses developing and deploying their own AI applications need to address a variety of vulnerabilities to ensure these systems are safe and secure for customers.”

“Cisco AI Defense is a single, end-to-end solution that helps your organization understand and mitigate risk on both the user and application levels. To accomplish this, it comprises four main components: AI Access, AI Cloud Visibility, AI Model & Application Validation, and AI Runtime Protection,” Sampath added.

Cisco says its AI Defense solution is expected to become available for enterprises in March.

Related: Biden Signs Executive Order Aimed at Growing AI Infrastructure in the US

Related: How to Eliminate “Shadow AI” in Software Development

Related: Trolley Problem, Safety Versus Security of Generative AI

Related: The Ghost of Christmas Past – AI’s Past, Present and Future