As hacking incidents exploiting artificial intelligence continue to mount, South Korea's cyber defense capabilities scored 58 out of 100, falling short of the global average of 64.
Cisco released its Relentless Defense Report containing the findings at a media roundtable on "Cybersecurity in the AI Era" held on Oct. 13 at its Cisco Korea office in Seoul's Gangnam district. The report is based on a survey of 8,000 security professionals across 30 countries.
Only 1% of Korean organizations qualified as "top-tier defenders," the category with the highest level of security capability, according to the report. That is far below the global average of 8%. By contrast, 80% of Korean organizations experienced a cybersecurity incident with a material impact on their business over the past year, lower than the global figure of 91%. More than one in three security incidents involved an AI-enabled attack.
The report identified internal friction as the main factor undermining defense capabilities. Roles and responsibilities are unclear, and departmental turf-guarding makes problems difficult to solve through the security team's efforts alone. While 41% of organizations globally said their security, networking and information technology teams share information and tools and collaborate as a single team, the figure was just 24% in Korea.
Cumbersome approval processes were also flagged. Responses saying security metrics are too technical to persuade management were similar in Korea at 40% and globally at 43%. Korea recorded a higher rate than the global average on spending more time collecting and connecting data from scattered systems than on responding to threats, at 44% versus 40% worldwide.
The survey also found that while AI adoption delivered substantial benefits, uptake has been slow. More than eight in 10 security practitioners said AI had improved their threat detection and response capabilities by at least 25%. By contrast, one-third of organizations had broadly deployed AI agents in security operations — one-fifth in Korea — and 54% were confident they could respond if an AI agent were compromised, compared with 37% in Korea.
For top-tier defenders, organizational capability mattered more than the scale of investment. Of the top-tier defenders, 83% operated as a single integrated security team sharing information and tools. The share applying consistent security controls to AI agents and non-human identities reached 86%, and 71% of those that expanded security investment reported a decline in incidents. In the remaining group, the corresponding figures were 37%, 37% and 39%.
Carl Solder, chief technology officer for Cisco Australia and New Zealand, said leading organizations are not simply using AI as a chatbot bolted onto the security operations center. "They are using the network itself — the fabric of interconnected devices through which data moves — as a sensor for the entire security stack and as a policy enforcement engine," he said.