[iNews24 Reporter Yurim Jeong] Naver on the 23rd began enforcing a policy that automatically disables comments on articles when the number of malicious comments passes a set threshold.
According to Naver, its AI-based detection system, Cleanbot, scans articles across all sections — including politics and elections — to identify malicious comments.
If the threshold is exceeded, Naver displays a notice saying, "Cleanbot has detected a large number of malicious comments and the commenting feature is not available," along with a banner promoting a healthier online environment. Because the system could be abused, Naver does not disclose the exact detection criteria.
Naver introduced the industry's first malicious-comment detection system in 2019 and has continued refining it to detect profanity, sexual or violent language, and hate speech as well as demeaning and discriminatory expressions. The company plans an AI-model upgrade later this month.
Kim Suhyang, a Naver executive, said, "We will continue to gather diverse input and develop the service so it can respond to rapidly changing forms of malicious comments."
Meanwhile, Naver also offers a memorial-comments feature for disaster, accident and obituary coverage to help prevent secondary harm to victims and the deceased while allowing readers to express sympathy and condolences. About 23 news outlets have adopted the feature so far, Naver says.
An analysis through April found that the article with the most memorial comments had a comments-to-page-views ratio about six times higher than other articles from the same outlet.
