Cisco "Growing Security Threats in the AI Era, Supported by 'AI Defense'"

Cisco Connect Korea 2025 Event Held
AI Security Solution 'AI Defense' Announced
Enhanced 'Webex AI Agent' Also Introduced

"In the process of introducing artificial intelligence (AI) into products or services, security should not be sacrificed for speed. Security must be a priority when building AI services.

Anand Ragavan, Vice President and Head of AI Products at Cisco, said this during a virtual attendance at the 'Cisco Connect Korea 2025' media roundtable held on the morning of the 3rd at the Grand InterContinental Seoul Parnas in Gangnam-gu, Seoul. He emphasized that AI security must be treated as an important aspect when applying AI to corporate tasks or systems.

Anand Ragavan, Vice President and Head of Cisco AI Products, is delivering a virtual presentation at the 'Cisco Connect Korea 2025' media roundtable event held on the morning of the 3rd at the Grand InterContinental Seoul Parnas in Gangnam-gu, Seoul. Photo by Cisco Korea

Cisco Connect Korea 2025 is considered Cisco Korea's largest annual event. Besides Vice President Anand, the media roundtable held prior to the event featured presentations by Choi Ji-hee, CEO of Cisco Korea, and David Coventry, Managing Director of Cisco Asia Pacific, Japan, and China (APJC) Collaboration division.

In his presentation, Vice President Anand introduced Cisco's AI security solution called 'AI Defense.' AI Defense is a solution that helps companies safely develop, deploy, and protect AI applications. AI Defense integrates with the security platform 'Cisco Security Cloud' to enhance security at the network level using AI-based technology.

The reason such AI security solutions are necessary is that AI applications and large language models (LLMs) operate simultaneously during AI service operations. Some services utilize multiple apps or LLMs, and during this process, attacks such as 'prompt injection,' where malicious commands are input, can occur. Since apps and LLMs operate organically within AI services, a security issue in one area can affect the entire service.

Cisco's AI Defense leverages AI's analytical and learning capabilities to detect and respond to threats in real time. At the same time, it strengthens the overall security of corporate AI service operations, according to Vice President Anand. He stated, "As AI is increasingly used in core areas across all organizations, three aspects have become very important: visibility into model usage, model validation before deployment in operational environments, and guardrails for model protection in operational environments." He emphasized, "AI Defense protects AI-based applications based on safety, security, and privacy guardrails."

Following Vice President Anand, Managing Director David Coventry introduced the features of 'Webex AI Agent,' an AI service available in Cisco's collaboration tool 'Webex.' This service combines AI with human agents to reduce customer response times and automate problem resolution. Director Coventry explained, "Webex AI Agent enables natural interactions across both voice and digital channels," adding, "It not only accurately understands customer intent but also integrates with existing systems."

Choi Ji-hee, CEO of Cisco Korea, is delivering a greeting at the 'Cisco Connect Korea 2025' media roundtable event held on the morning of the 3rd at the Grand InterContinental Seoul Parnas in Gangnam-gu, Seoul. Provided by Cisco Korea

In her welcoming remarks, CEO Choi stated, "Cisco, which has led the networking field for the past 40 years, is transforming into a company driving security, AI, and data innovation," and added, "Cisco will support corporate customers' AI transformation and digital resilience with networking, security, collaboration, and visibility solutions that have significantly enhanced AI capabilities."

During the subsequent Q&A session, there was a forecast that the importance of AI security will increase this year. Vice President Anand diagnosed, "Last year, customers were in the early stages of considering AI security issues, but this year, the discourse has shifted to how AI can be used safely." CEO Choi also said, "Although many domestic customers do not yet seem to be deeply concerned about AI security, security issues will be addressed more seriously once AI agent services are introduced."