[DiaryPoint = Reporter Kim Jeong-hun] South Korea’s Ministry of Science and ICT and the AI Safety Institute appointed 80 members to a public monitoring group called “AI Safety Guardians” on November 12. Selected from 138 applicants who applied during an open recruitment drive in August and September, the members will identify and report AI errors and risks through August 2027.
Participants range from high school students to people in their 60s and older, and include office workers, homemakers, freelancers and students. The group is made up of members selected through open recruitment. Other members of the public must apply through a separate recruitment notice and selection process.
The monitors will look for problems in four areas: AI-generated factual or numerical errors; bias against particular groups; security vulnerabilities; and defects such as failing to answer or repeating the same content.
Reports will be checked for validity based on accuracy, risk level, reproducibility and potential usefulness. Verified cases will serve as source material for AI safety evaluations and risk research. The ministry and institute also plan to share cases with relevant companies.
An earlier example of ongoing public monitoring was the “Real AI Challenge: No Misinformation” campaign, launched in December 2025. Of 625 reports submitted over about 30 days, 464 were confirmed as valid error cases.
At the launch ceremony, Eom Yeol, head of the ministry’s AI Policy Office, said, “As AI has become part of daily life for people of all ages, creating an environment where everyone can use it safely is something we must do together.” After the ceremony, the appointed monitors received training on how to identify errors and the criteria for submitting reports.