Increasing network visibility enables quick mitigation of security threats. This is the primary premise of the development of Cisco AI Defense as more and more companies are deploying artificial intelligence (AI) applications.
“AI safety and security is probably one of the biggest challenges we all need to address,” Dave West, president for Asia Pacific, Japan, and Greater China (APJC) at Cisco, said in a virtual media briefing. “AI is evolving and changing very quickly.”
West also noted the increasing adoption of AI among businesses and organizations. However, in Cisco’s 2024 AI Readiness Index, only 35% of those surveyed in the Philippines feel fully equipped to detect and prevent unauthorized tampering with AI.
“In the very near future, there will be only two types of companies: those that are AI-forward and those that are irrelevant,” West pointed out.
The complexity of AI deployment includes applications being multi-model and multi-cloud. Add to that the different vendors for different solutions making it difficult to navigate operations.
Cisco developed AI Defense to help businesses use AI safely by solving two main challenges: building secure AI tools and protecting access to them.
“What Cisco AI Defense does is it does two things. The first thing it does is it basically protects users and employees that are using AI applications inside of your organizations,” said DJ Sampath, head of AI Software and Platform at Cisco.
Unified security measures across AI platforms
AI Defense simplifies this by providing developers with a unified set of safety and security measures that work across all AI platforms. It helps developers move quickly while ensuring their systems are protected from attacks and behave as intended. AI Defense also helps security teams discover all AI tools in use, whether officially approved or not, across both public and private clouds. It automatically tests AI models to detect harmful or unexpected behaviors and recommends fixes to address these issues. It also provides continuous protection by monitoring for risks like data leaks, hacking attempts, or malicious inputs that could disrupt the system.
AI Defense also focuses on protecting access to AI tools as employees increasingly use them to improve productivity. Security teams gain full visibility into the AI applications employees are using, ensuring that unapproved tools are identified and restricted. Policies can be enforced to block access to these tools, and sensitive company data is continuously safeguarded from leaks and threats while ensuring compliance with industry regulations.
Machine learning
Unlike safety measures built into individual AI models, Cisco’s AI Defense promises to deliver consistent controls across multiple models. It is powered by Cisco’s proprietary machine learning systems, which adapt to new threats using intelligence from Cisco Talos. The system integrates seamlessly with existing data workflows, offering “unmatched visibility” and control through Cisco’s Security Cloud platform.
By securing AI systems at the network level, Cisco aims to provide a level of accuracy and reliability that sets it apart, while contributing to the development of industry standards for AI security.