Singapore courts may lean towards fair use as a defense for machine learning under certain circumstances, despite the absence of established case law. The Singapore Copyright Act’s fair use provision could potentially be invoked to protect machine learning activities within specific parameters.
In a recent case, Global Yellow Pages (GYP) v Promedia, the Court of Appeal in Singapore allowed the defense of fair use under section 35 of the Copyright Act. The case involved the Defendant photocopying and scanning the Plaintiff’s business listings for data access in creating their own listings. The Court considered key factors in applying fair use, including the purpose of the dealing, the nature of the work, the amount used, and the impact on the potential market.
Applying this reasoning to machine learning, where crawling fact-based websites is involved, fair use defense could be effective. The nature of the copyrighted material, prevention measures against infringement, and adherence to terms of service play a crucial role in determining fair use applicability. Implementing output filtering techniques, clear guidelines, and user responsibilities can strengthen the case for fair use in machine learning.
While the legal landscape around AI-generated content is evolving, measures like output filtering and user compliance guidelines are essential to mitigate copyright infringement risks in machine learning applications. The effectiveness of such measures and adherence to fair use factors under the Singapore Copyright Act are vital considerations for courts in determining the legality of machine learning practices. Similar considerations apply to neighboring countries like Malaysia, sharing the same fair use factors in their Copyright Act.
