Jakarta. Indonesia is studying the creation of a collective management organization that would collect royalties from artificial intelligence (AI) companies using copyrighted content to train AI systems.

Director General of Intellectual Property at the Legal Affairs Ministry, Hermansyah Siregar, said the proposed Collective Management Organization (LMK) would simplify royalty payments for copyright holders whose works are used by AI firms for automated data training.

“If companies have to seek permission individually from every copyright holder, it would be difficult because the dynamics of AI development move very quickly,” Hermansyah told lawmakers during a hearing with Commission XIII of the House of Representatives in Jakarta on Tuesday.

He explained that AI companies train their systems using large amounts of existing content collected from the internet, books, music, videos, and other creative works. Such use, he said, should require authorization from copyright owners.

Hermansyah said that certain exemptions may apply for educational, health, or research purposes. However, AI companies using copyrighted material outside those categories would still be required to obtain licenses and pay royalties.

“For example, songs, artwork, and other creative content should be licensed, and creators must receive royalties,” he said after the parliamentary meeting.

According to Hermansyah, it would be impractical for AI companies to compensate creators individually because AI systems often crawl and process data from millions of creators simultaneously.

Under the proposed system, AI companies would instead pay royalties to a single collective management organization, which would then distribute payments to creators.

“The LMK would later distribute the royalties to the creators,” he said.

The proposal is expected to be included in revisions to Indonesia’s 2014 Copyright Law, which the government is currently reviewing.

Hermansyah said Indonesia is still refining the concept and studying similar practices in other jurisdictions, including the United States, the European Union, and China, where debates over AI training data and copyright protections have intensified in recent years.

“Because this is a new issue, we need comparative studies from countries that are more mature in handling these matters,” he said.