On August 26, Jongheon Yang, Head of S2W’s Offensive Research Team, spoke at the International Symposium on Cybercrime Response 2025 (ISCR 2025) in Seoul, sharing insights on emerging security risks in the age of generative AI.
Yang presented eight major threats exploiting LLM vulnerabilities and outlined practical countermeasures. He also reviewed notable generative AI threat cases from this year, stressing the importance of strong data protection policies, fine-grained access control, and continuous monitoring.
S2W is addressing these challenges through Security Guardrails, a defense framework that ensures safety and reliability throughout the development, deployment, and operation of LLMs. These guardrails are integrated into SAIP, S2W’s industrial generative AI platform, enabling effective protection against risks that may arise in generative AI services.
The company also applies ontology-based knowledge graph technology to map relationships between malware, vulnerabilities, and threat actors. This approach helps organizations build effective defense strategies against increasingly sophisticated AI-driven threats.
Learn more about SAIP, the industrial generative AI platform:
🤖 https://s2w.inc/en/product/pd01
Learn more about QUAXAR, the cyber threat intelligence platform: