“Artificial intelligence (AI) also has security issues and there are attackers who exploit AI for ha..

“Artificial intelligence (AI) also has security issues and there are attackers who exploit AI for hacking, but some hackers are using AI in a society-friendly way. I want to be a hacker who spreads positive influences.” (Winning team Blue Water leader Jasra)

With the rise of Generative AI technology, security threats from AI are increasing, such as generating sophisticated phishing emails with AI. Meanwhile, the white hacker camp that prevents this is also actively using AI to respond. Leading hackers who participated in the International Hacking Defense Competition ‘Codegate 2025’, which determines the world’s best white hackers, were actively using Generative AI tools such as ChatGPT and Claude to solve problems one by one.

Codegate is an event hosted by the Ministry of Science and ICT and organized by Mail Business, the Codegate Security Forum, and the Korea Internet & Security Agency (KISA). It is one of the world’s top three international hacking defense competitions, and this year, hackers from all over the world competed for the total prize money of 71 million won.

The final match was held at the COEX Grand Ballroom in Seoul on the 11th with more than 100 athletes participating. 2,778 people from 66 countries participated in the preliminary round held in March this year, and 40 teams that passed the fierce competition made it to the final round. In the final, which was divided into 20 teams each into the general division and the junior division, the winner of last year’s championship and the combined Korean, Indian, Chinese and Canadian teams “Blue Water.”

The Blue Water Team showed overwhelming skills in major tasks such as attacking vulnerabilities in the Windows operating system (OS) and attacking the setting of a new technology “MCP” in AI, recording its second consecutive competition with a gap of nearly 3,000 points with the second-place team. Bluewater team leader Jasra said, “We have overcome complex problems with the help of the Large Language Model (LLM).” The second and third places in the general division went to the Korean teams “CyGon” and “The Duck.”

The topic of Codegate 2025, which was held under the theme of “Universal Artificial Intelligence (AGI) and Cybersecurity,” was AI and security threats. In the junior finals under the age of 19, there was also a question of attacking and bypassing LLM’s vulnerabilities.

In the junior final, three students in their third year of Sunlin Internet High School swept the first to third places. Student Park Jin-woo, who won the championship, said, “I want to become a researcher in the security field after graduation,” adding, “I want to help Korea become a security powerhouse.” Kim Min-ki, a teacher at Sunlin Internet High School who guided security prospects, said, “We will contribute to creating a social atmosphere in which the security field is in the spotlight as a student career path.”

Jang Miao, general manager of Huawei’s cybersecurity trust building, appeared in the keynote speech that day, drawing attention. General Manager Jang emphasized security governance in the AI era and said, “As human life extends to the digital world under AI influence, AI security is expected to become more important.” Jang explained that it is important not only for companies but also for the government to share security responsibilities together.

Codegate was launched in 2008 by Hancom Group Chairman Kim Sang-chul to train security talent, marking its 17th anniversary this year. Jang Seung-joon, vice chairman of Maekyung Media Group, co-organizer of Codegate, said in a congratulatory speech, “AI services will be implemented properly only when AI security is supported as much as AI technology.”

[JUNG HOJUN / Reporter Kim Daegi]