PETALING JAYA: While proposed reforms to the Copyright Act 1987 are long overdue, industry players say key issues, particularly the use of copyrighted works to train artificial intelligence (AI) systems, remain unresolved.
Intellectual property lawyer Sabrina Mohamed Hashim said amendments are necessary as the Act predates AI systems, large-scale data mining and modern digital distribution models.
ALSO READ: Better copyright protections in store
“As AI becomes more widely adopted, the lack of clear rules creates uncertainty for creators, rights holders, businesses and developers, particularly on issues of ownership, consent and compensation,” she said.
Karyawan president Datuk Freddie Fernandez said the organisation supports the proposed amendments.
He said reforms must be backed by a government-controlled centralised national music rights library, royalty collection and distribution platform.
“Without such a system, AI protection will remain difficult to enforce, royalty disputes will continue and artistes may still be deprived of fair income despite legal reform,” he said.
Fernandez said the Act should explicitly prohibit copyrighted music, lyrics, sound recordings, performances and artistes’ voices from being used to train AI systems without consent or a lawful licensing framework.
Universiti Sultan Zainal Abidin professor Dr Jeong Chun Phuoc said the Copyright Act does not specify AI model training, machine learning, text and data mining (TDM) or the use of copyrighted works in AI datasets.
“Malaysia should introduce a dedicated legal framework covering AI training, licensing, transparency, remuneration and TDM to balance creators’ rights with responsible AI innovation.”
