Japan Plans AI Pilot Program to Fight Manga & Anime Piracy

Home > Anti-Piracy >

Japan's Cultural Affairs Agency is looking towards AI to strengthen the country's battle against rampant online piracy. The pilot program proposed by the agency would see culturally significant manga and anime content protected by automated image and text recognition systems, with less reliance placed on resource-intensive human intervention.

japan-ai-protection With the arms race between pirate sites and rightsholders showing no obvious signs of cooling down, the online piracy landscape is becoming increasingly complex.

The existing toolbox of legal measures, such as action against domain names, site-blocking, search engine penalties, even whole-site deindexing, has led to the emergence of a new breed of shape-shifting, measure-evading, pirate sites.

Seemingly able to rebrand almost at will, it’s not uncommon for sites to reappear with new names and a new coat of paint, without any significant damage to existing traffic. Meanwhile, new sites and new brands, sporting multiple new domains, continuously vie for attention, as the illegal content continues to flow.

Human Moderators Can “Barely Keep Up”

In a statement issued this week, Japan’s Cultural Affairs Agency said the proliferation of illicit content online costs the country’s manga and anime industries billions of dollars each year in lost revenue.

Recent figures discussed as part of Japan’s comprehensive plan to combat infringement estimate the damage as somewhere between 1.9 trillion and 2.2 trillion yen (US$12.6bn and US$14.6bn); enough to “significantly increase Japan’s GDP” and “potentially eliminate” the trade deficit in content-related services, documents suggest.

The scale of the problem is well understood, but with an estimated 1,000 sites offering culturally significant manga and anime content for free, monitoring for the availability of pirated content online is a resource intensive task. Human moderators can now “barely keep up” due to the proliferation of illicit content online.

“Copyright holders spend a significant amount of human resources trying to manually detect pirated content online,” said Momii Keiko, a director at the Cultural Affairs Agency’s Copyright Division.

To level the playing field, Japan is looking towards a future where detection of illicit content will be less reliant on human intervention.

AI-Powered Pirated Content Detection System

To improve pirated content detection rates, Japan’s Cultural Affairs Agency has filed a supplementary budget request of 300 million yen (US$2 million) to fund a pilot program that will use artificial intelligence to detect pirated content online. The aim is to detect content using image and text recognition systems.

The proposed pilot is reportedly inspired by a similar program in South Korea. Announced mid-2023 by the Ministry of Culture, Sports and Tourism, the program aimed to overcome the limitations of manual work based on human resources, by automatically detecting the emergence of new, cloned, or substitute pirate sites.

At least initially, the pilot program in Japan will focus on pirated manga and anime. If successful it could also be expanded to other content such as films, TV shows, music, and content from Japan’s wider publishing industries.

Success of Automated Systems Totally Depends on Accuracy

The success of automated detection systems lies in their accuracy; nobody benefits if detections lead to false allegations of infringement, or if illegal content slips through the net. These concerns appear to be front and center in Japan as rightsholders continue to tackle pirate sites.

TorrentFreak was recently able to review a transcript of a discussion which focused on international enforcement efforts spearheaded by the Content Overseas Distribution Association (CODA).

While highlighting anti-piracy successes and challenges in various countries, emphasis was placed on the importance of balancing anti-piracy measures with freedom of expression and the protection of communications privacy.

Sponsors

Popular Posts

From 2 Years ago…