Law Firm Emphasizes Need to Mitigate Risks of AI Agents
October 18, 2026 · 2 min read
Also published in हिन्दी
As the utilization of AI agents expands, the management of security within corporate internal systems has emerged as a new challenge. There is a need for a system that continuously checks activity logs and restricts access rights, given the possibility that AI agents may act unexpectedly or leak personal information externally.
On October 15, a seminar titled 'Personal Information Protection and Incident Response Strategies for Online Shopping Companies in the AI Era' was held at the Sejong D Tower in Seoul, organized by a law firm in conjunction with the Korea Online Shopping Association. This event highlighted the increasing cyber attacks utilizing AI and the tightening regulations on personal information protection.
Professor Choi Dae-seon from Soongsil University explained that advancements in AI technology have reduced the time and cost required for cyber attacks, thus making information access and rights management for AI agents used internally by companies increasingly important.
Professor Choi stated, "AI agents can potentially act outside of corporate directives and expose personal information externally. While restricting access rights or requiring prior human approval may help, it is difficult to achieve perfect control, which necessitates a system to continuously record and check the actual activities of AI agents."
Choi Kwang-hee, a senior adviser, emphasized the importance of preserving evidence immediately after an incident occurs and ensuring timely reporting and notification in accordance with legal deadlines. He explained that having a system in place for training responses based on real incidents can minimize the spread of damage.
Ahn Jeong-ho, a lawyer at the firm, discussed the enhanced preventive obligations of companies due to the amendments to the Personal Information Protection Act and the Information and Communications Network Act, emphasizing that personal information protection and information security have become responsibilities that must be managed at both the operational and executive levels.