Naver Deletes Nearly 40 Million Harmful Posts

Korea's leading online portal Naver deleted nearly 40 million harmful posts in second half of last year, most with the help of artificial intelligence.

From last July to December, a total of 38.53 million posts were removed for violating Naver's policies, and spam or advertising accounted for over 90 percent. The remaining posts included foul language, discrimination, hate, content harmful to minors, and illegal information.

AI played a significant role in automatically filtering and removing sexually explicit content, with 98 percent of porn videos being screened and deleted in real time. In the case of news, more than 5 million comments containing hate speech, discriminatory language, insults and defamation were removed before they could even be posted thanks to algorithms that detect them.

Naver said any harmful posts and malicious comments that AI fails to detect are then flagged by users, and a dedicated team reviews and removes them.

Naver also received 95,839 requests to delete content that violated rights such as privacy and copyright over the period. More than 80 percent was potentially defamatory.