“AI Is Already Part of Workflows”: Security Firms Offer Enterprise 'Security Control' Solutions

As generative AI has become a basic element of the work environment, some employees are increasingly using AI agents that go beyond answering questions and also carry out tasks. However, it remains unclear whether security risks are being addressed at a pace comparable to AI transformation (AX).

In response to this trend, security companies are launching new solutions aimed at solving these problems and stepping up efforts to capture the market. At the same time, venues are being prepared to take a multifaceted look at how AI is changing cybersecurity and to discuss the challenges ahead for companies and institutions.

Computerworld/IT DAILY will hold the '2026 Information Security Solution Conference' on Sept. 9 at the Orche Hall on the 5th floor of L Tower in Yangjae-dong, Seoul. The event theme is 'AI for Security, Security for AI,' and it is expected to serve as a forum for discussing the changing cybersecurity landscape driven by AI and the tasks ahead for companies and institutions.

The Bank of Korea released a report titled 'Rapid Spread of AI and Productivity Effects' last August, saying that 63.5% of workers had experience using generative AI and that the generative AI utilization rate for work purposes stood at 51.8%. It also found that Korea's work-related usage rate was about twice that of the U.S. at 26.5%, and that the average time spent using generative AI was 5 to 7 hours per week.

Amid this spread, domestic companies were assessed as being at a point where it is difficult to find someone who has not used generative AI, and the perception that AX is lagging in industries with a high share of closed networks has also weakened. The Software Policy and Research Institute (SPRi) released a '2025 Research on AI Adoption in the Public Sector' in April, and said that 667 AI adoption contracts for public-sector work efficiency in 2024 marked a roughly 6-fold increase from 98 public AI adoption contracts in 2015.

Along with rising interest in AX, the use of generative AI is also increasing, but concerns are being raised about how well companies and institutions are prepared for the security risks behind AX. According to Cisco's '2025 Data Privacy Benchmark Study,' 64% of respondents were concerned about the possibility of information leaks stemming from generative AI. In the same study, 46% of respondents said they had entered employee names or information into generative AI tools, and 42% said they had entered corporate confidential information into generative AI tools, highlighting both risk awareness and the lack of clear countermeasures.

In this situation, some argue that restricting generative AI solely for security reasons is inappropriate. Building an internal Private AI can ease concerns about data leaks, but achieving satisfactory performance with Private AI alone is difficult. Accordingly, establishing how to control risks while allowing both internal and external AI services is presented as a corporate challenge in the AX process going forward.

SoftCamp stressed that, with the industry entering the AI agent era, security is not adequately addressed by a simple allow-or-block dichotomy. It also said that defining the targets and scope of AI delegation is key. SoftCamp proposed a solution in which permissions, data, and execution scope are policy-based limits on the premise of maintaining AI utilization.

SoftCamp presented 'AI Security Suite' as its solution and organized its AI security enhancement strategy around four control axes. Control axis 1 is identity security 'Shield ID,' control axis 2 is external service isolation 'Shield Gate,' control axis 3 is centralized MCP and API control 'Shield AI Gateway,' and control axis 4 is AI agent identification and permission management 'Shield Agent Sandbox.'

WINS Technet offers 'SNIPER AIVAX,' a technology designed specifically for AI-based Data Loss Prevention (DLP). The technology detects sensitive data such as personal information, corporate secrets, and technical information in prompts and blocks it before it is sent outside, while also protecting against unintended information leaks and malicious prompt Injection attacks.

'SNIPER AIVAX' is provided in the form of network equipment. Companies can either install agents tailored to their internal environment and choose proxy-based controls, or connect to general-purpose AI services using an API key and then apply controls.

The '2025 Information Security Solution Conference' was held on Sept. 4 last year at L Tower in Yangjae-dong, Seoul. Photo source: Computerworld/IT DAILY.

On the 9th, the '2026 Information Security Solution Conference' will be held. The keynote will be delivered by Kim Chang-hoon, professor in the Department of Information Security at Daegu University, who will speak on the topic of 'AI Data Center (AIDC) Cybersecurity Framework.'

Next, Choi Young-cheol, CEO of SGA, will present a server security and microsegmentation-based zero-trust strategy for the AI era, and Han Tae-dong, principal manager at AhnLab, will introduce a data security strategy for the prompt-centric generative AI era. Kim Joo-seon, director at WINS Technet, will present an AI-DLP application plan in response to the transition to the National Network Security System (N2SF), and Kim Dong-hoon, vice president of SoftCamp's consulting division, will introduce 'AI Security Suite' for AI agent use. Kim Ki-woong, security strategy response team leader in KB Kookmin Bank's Information Security Department, will present 'Financial Sector AI Attack Simulations and Response Strategies.'

Event participation conditions were also announced. Attendance at the '2026 Information Security Solution Conference' is free for preregistered participants, and preregistration is available on the IT DAILY website.

Source: IT DAILY · Kim Ho-jun
Original: https://www.itdaily.kr/news/articleView.html?idxno=241130

References

• https://www.itdaily.kr/news/articleView.html?idxno=241130