KUALA LUMPUR (Oct 6): Malaysia is studying the possibility of introducing provisions that allow copyright owners to opt out of having their works used for data mining, including for training artificial intelligence (AI) systems.

Intellectual Property Corporation of Malaysia (MyIPO) director general Yusnieza Syarmila Yusoff said the opt-out mechanism is among the key elements being considered in reviewing the nation's copyright legal framework to keep pace with rapid technological advancement.

"We are exploring the possibility of introducing provisions related to data mining that also give rights holders the choice to opt out, namely, not allowing their works to be used," she said when serving as a panellist at a media forum here on Tuesday.

Yusnieza said MyIPO has held engagement sessions with stakeholders to assess the implications of tech companies using copyrighted works as training data to develop AI systems and generate new content.

She said AI-related copyright issues encompass two main aspects, namely the use of works as input or training data, and the ownership status of output generated by the technology.

"AI involves two main issues, namely input regarding how training data is used by the technology, and secondly, the output issue regarding the ownership status of generated results," she said.

In view of this, the ongoing review also covers the scope of permitted use of works for data mining purposes, including detailing the types, purposes, and volume limits of content that can be accessed.

She noted that any new legal framework must strike a balance between protecting copyright owners and fostering an environment conducive to technology, trade, and investment.

"We do not want laws to be overly restrictive to the point of stifling trade, investment, and technological growth. The framework established must be user-friendly, IP-friendly, and balanced," she said.

Yusnieza added that studies into every provision, exception, and limitation under copyright law were ongoing through forums and discussions with stakeholders before any approach is finalised.

Uploaded by Tham Yek Lee