Naver Opens Technology for Real-Time Pornographic Content Filtering System

[Asia Economy Reporter Seungjin Lee] Naver announced on the 16th that it has revamped its real-time obscene content filtering system 'XAI 2.0' as 'Clova GreenEye' and will open the technology to external users.

'XAI' is a real-time obscene image filtering technology launched by Naver in 2017. It learns from millions of images accumulated on Naver, and when inappropriate images or videos (obscene content) are uploaded to Naver, AI technology detects them in real time and blocks their exposure in search results.

Naver has utilized 'XAI' across various services such as Cafes, Blogs, and Knowledge iN to block obscene content and provide safer services to users. In 2021, the system was upgraded to 'XAI 2.0', which classified content into categories such as normal, obscene, adult, and suggestive. It also applied various advanced image recognition research results from Naver, including 'RexNet', 'CutMix', and 'AdamP', improving accuracy up to 99.5%.

'Clova GreenEye' can be applied for on the Naver Cloud Platform website, and during the beta service phase, it is provided only to approved users after application. Naver expects that with the launch of 'Clova GreenEye', startups and small to medium-sized enterprises without technological capabilities will be able to easily automate harmful content detection and minimize manual review, thereby providing stable content services.

Lee Jung-kyu, the lead responsible for Naver's service operations, stated, “By opening the obscene content filtering technology that has contributed to user protection within Naver for over five years to external companies, we aim to contribute to creating a healthier internet ecosystem.” He added, “We will continue to advance AI models, update data, and expand the scope of harmful content detection.”

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