Japan Govt. Eyes AI-Based System for Detecting Pirate Manga Sites, with ¥100 Mil. Earmarked in Draft Supplementary Budget

Yomiuri Shimbun file photo
The Cultural Affairs Agency

The government aims to realize the use of a system utilizing AI that automatically detects websites displaying manga without permission as early as in fiscal 2026.

The envisioned system allows AI to train on image and text data taken from manga provided by publishers to detect pirate sites and unauthorized social media posts. The system has been in trial operation since fiscal 2024 targeting domestically hosted pirate sites.

According to the ABJ (Authorized Books of Japan), an antipiracy group made up of publishers and other entities, estimated losses from people accessing manga and other pirated works for free stood at about ¥700 billion in June alone.

Currently, publishers have the burdensome task of manually checking websites and social media posts to find pirated content and request website operators take it down. It is said that not all pirate websites can be detected and the volume of pirated content is too large to deal with through data deletion requests. This has made creating effective countermeasures an important task.

The Cultural Affairs Agency plans to demonstrate the system targeting 10 countries home to many websites hosting pirated content, to expand detection coverage and improve accuracy before practical implementation.

It will also study a system that automatically requests the deletion of pirated content and issues warnings to pirate site operators on behalf of publishers.

The agency earmarked ¥100 million in the draft fiscal 2025 supplementary budget.