- Cisco Unveils New Tool to Ensure Security of Enterprise AI Applications
- Cisco AI Defense Seeks to Manage Risks Associated with Developing and Deploying Secure AI Applications
- AI Defense is self-optimizing and will improve over time, Cisco says
Cisco has unveiled a new solution aimed at helping businesses build, deploy and secure applications based on artificial intelligence (AI).
The company says Cisco AI Defense addresses two key risks: the development and deployment of secure AI applications.
According to Cisco, companies are expected to develop “hundreds, if not thousands” of AI applications. They can’t afford separate sets of AI security and safety guardrails for each application: they need a unified approach.
Multiple capabilities
Cisco says AI Defense capabilities include AI discovery (giving security teams insight into who is building applications and what training sources they use), model validation (minimizing the risk that tuning the model becomes toxic through automated testing), runtime security (continuous validation that protects against potential safety and security threats such as rapid injection), and securing access to AI applications.
The company says AI Defense “self-optimizes,” providing controls for a multi-model world. Additionally, it integrates with existing data feeds and is integrated with Security Cloud, Cisco’s cross-domain security platform.
The new tool will be available to businesses in March 2025.
“Business and technology leaders cannot afford to sacrifice security for speed when adopting AI,” said Jeetu Patel, executive vice president and chief product officer at Cisco. “In a dynamic landscape where competition is fierce, speed decides the winners. Integrated into the network fabric, Cisco AI Defense combines the unique ability to detect and protect against threats when developing and accessing AI applications, without compromise.
