What concerns do Malaysian music publishers have about AI training data?

Question

Accepted Answer

Malaysian Music Publishers Express Concerns Over Unauthorized AI Training Data Use

Malaysian music publishers are closely monitoring the potential use of locally produced songs in generative AI model training without proper authorization. This concern stems from the discovery of extensive song datasets, identified by The Atlantic, which are used within the AI development community. A search of these datasets revealed a significant number of Malaysian musical works, prompting publishers to assess their rights and advocate for clearer regulations.
Discovery of Malaysian Songs in AI Datasets

An investigation by The Atlantic journalist Alex Reisner identified four large datasets comprising over 21 million songs that are reportedly used in AI development. Following this, The Atlantic launched The AI Watchdog, a search engine designed to identify works present in these datasets. A review of this tool indicated the inclusion of at least 1,000 songs by Malaysian artists, spanning various genres and levels of fame.
Prominent Malaysian Artists Found in Datasets

The datasets contain a notable number of songs from well-known Malaysian musicians. For instance, Datuk Seri Siti Nurhaliza has approximately 106 songs listed, while the extensive catalogue of the late Tan Sri P. Ramlee features around 134 songs. Contemporary artists like Yuna have 151 songs across the identified datasets, and the rock band Butterfingers has 73 songs. Other artists whose works were found include Carburetor Dung, Fish Leong, Alleycats, Jess Lee, Spooky Wet Dreams, Hujan, Ziana Zain, and Dayang Nurfaizah.
Publishers' Stance and Calls for Transparency

Taja Archive Sdn Bhd Malaysia director Kimmy Khir stated that music publishers are actively observing the situation to safeguard the rights of songwriters and copyright holders. While acknowledging that presence in a dataset does not definitively confirm AI training use, publishers are prepared to take action if credible evidence emerges of unauthorized use impacting rights holders. Taja Archive, representing approximately 900 songwriters and administering nearly 20,000 songs, advocates for increased transparency, clearer licensing practices, and direct engagement between AI developers and rights holders.
Advocacy for a Robust Legal Framework

The Malaysia Music Rights Alliance (MMRA), a joint initiative by the Music Publishers Association of Malaysia (MPA) and the Recording Industry Malaysia (RIM), is actively campaigning for a comprehensive legal framework to govern AI's role in music. This alliance has communicated with Malaysian authorities, offering its expertise to help establish regulations that protect the nation's creative economy from potential negative impacts of unchecked technological advancement.

Reading history

Hide →

This history is cleared when the page is closed.

What concerns do Malaysian music publishers have about AI training data?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends