Negative Comments Violating Naver Regulations Decreased by 63%

[Asia Economy Reporter Bu Aeri] Naver announced on the 20th that the number of malicious comments deleted due to policy violations decreased by 63.3% in June compared to the beginning of the year.

During the same period, the number of "dislike" clicks decreased by 21.5%, and reports decreased by 53.6%.

Naver explained, "The 'comment history disclosure' implemented in March contributed to the reduction of malicious comment writing, while the 'blocking comments written by specific individuals' in April and the 'Cleanbot upgrade' in June significantly helped reduce the exposure of malicious comments."

The number of comments in June compared to the beginning of the year slightly decreased by 0.7%, while the number of authors increased by 8%. The number of comments deleted by the authors themselves decreased by 20.8%, and the frequency of comments per person decreased by 8%.

Naver analyzed, "There were concerns that disclosing comment history and implementing identity verification would shrink the comment space, but on the contrary, more users are participating and leaving more diverse voices more carefully."

As the collection of comments written by users was made public, visits to the "comment collection page" increased fourfold. It was also recorded that cases of users entering articles through the list at the bottom of the comment collection page increased by 22.5%.

The AI comment management system "Cleanbot 2.0," introduced on the 19th of last month, expanded the criteria for judging malicious comments from "profanity words" to "sentence context." Naver explained that Cleanbot 2.0 has an accuracy rate of 95%. Compared to before its introduction, the number of detected malicious comments doubled at most, and the number of malicious comment reports decreased by 19%.

Naver added, "In addition to improving the performance of AI Cleanbot, Naver will continue various policy and technical attempts to solve the problem of malicious comments."

This content was produced with the assistance of AI translation services.