Singapore authors raise copyright concerns over IMDA’s AI model

Photo credit: bee32 / 123RF

Singapore’s Infocomm Media Development Authority (IMDA) is in the throes of building a Southeast Asia-focused large language model. But the city-state’s literary circles worry that their copyrighted work might be used to train this model without consent, the Straits Times reported.

The IMDA has been working with local literary nonprofit firm Sing Lit Station to gather input in a survey, but writers say the research has a short timeframe and lacks clarity around compensation.

“If implemented without due consideration and safeguards, AI software will start assimilating material that would otherwise be copyrighted and adversely impact the livelihood of existing authors and publishers,” said Ng Kah Gay, editor at independent publisher Ethos Books.

These worries mirror those of the New York Times, which previously served a lawsuit to ChatGPT maker OpenAI and Microsoft for copyright infringement. The publication claimed that millions of its articles were used to train AI models that are being used to compete with the major publisher.

OpenAI has since countersued, saying that the New York Times hacked into its systems to gain evidence.

Singapore’s Southeast Asia-focused LLM, called The National Multimodal LLM Program, intends to understand the city-state’s unique language traits and multilingual context. Along with Singapore’s Sea-lion model, IMDA said these efforts aim to address AI’s limitations on regional context and languages.

For now, the IMDA has extended the survey deadline from April 7 to April 15, and it will remain online indefinitely.

See also: Mapping the key genAI chip and components makers in SEA