Singapore's MERaLiON, Southeast Asia's empathetic Large Language Model, Breaks New Ground with Multilingual Processing and Emotional Intelligence

Korea Herald

5 days ago

SINGAPORE, May 29, 2025 /PRNewswire/ -- At ATxSummit 2025, the flagship event of Asia Tech x Singapore (ATxSG), Mrs Josephine Teo, Minister for Digital Development and Information (MDDI), unveiled updates to Singapore's Large Language Model - Multimodal Empathetic Reasoning and Learning in One Network (MERaLiON) - and launched the MERaLiON Consortium, a significant boost to Southeast Asia's AI capabilities. She also announced new global initiatives and collaborations which strengthen Singapore's AI ecosystem and accelerate responsible AI adoption, putting Singapore at the forefront of efforts to operationalise AI safety.
MERaLiON breaks new ground with enhancements and new Global Partnerships
MERaLiON, a large language model developed by A*STAR Institute for Infocomm Research (A*STAR I 2 R) and supported by IMDA, breaks new ground in regional AI capabilities with enhanced multilingual processing and emotional intelligence. The model now handles Malay, Tamil, Thai, Bahasa Indonesia, and Vietnamese on top of English, Mandarin and Singlish, with advanced code-switching abilities and emotion recognition features. These improvements enable more intuitive and culturally aware AI applications in Southeast Asia, which could be applied to customer service, social service, and marketing.
To accelerate adoption, IMDA and A*STAR I 2 R have launched the MERaLiON consortium, to bring together local and global industry players and research and development (R&D) institutions like HTX, MOH Office for Healthcare Transformation (MOHT), NCS, National Supercomputing Centre (NSCC), SPH Media and ST Engineering with technology companies like Axiom IT Solutions, BytePlus, CommonTown, DBS, Grab and Microsoft Singapore. The consortium will focus on developing practical AI applications, from multilingual customer support to health and emotional insight detection and agentic decision-making systems.
Since its release in December 2024, MERaLiON's first version has garnered over 90,000 downloads globally, attracting users from corporate research labs, media service providers, startups, and academics.
Advancing global conversations for a trusted AI ecosystem
In the area of AI safety research, Singapore was the site of global cooperation in identifying and prioritising research domains through the "The Singapore Consensus on Global AI Safety Research Priorities" (The Singapore Consensus). This Consensus is a living document that will serve as the foundation for the ATxSG Government-to-Government (G2G) Ministerial Roundtable on Digital Trust which seeks to bridge science to policy, and translate technical research into practical policies, by facilitating meaningful conversations between AI scientists and policymakers.
Singapore's enhanced AI Verify Testing Framework addresses both Generative AI (Gen AI) and traditional AI risks. In continued collaboration with the US National Institute of Standards and Technology (NIST), we developed a crosswalk to map the enhanced Framework with the NIST AI Risk Management Framework. This reaffirms the alignment between our two countries' AI governance frameworks and reflects our shared commitment to cooperation.
Together, these initiatives underscore Singapore's commitment to building foundational infrastructure for AI safety development and testing that is grounded in scientific evidence to co-develop implementable, interoperable frameworks for responsible AI.
Strengthening global digital trust partnerships
Singapore's AI Safety Institute signed a Joint Statement with France's AI Safety Institute to kickstart collaboration and cooperation on AI safety and cooperation on 28 May 2025. The Joint statement was signed between MDDI, Mrs Josephine Teo, and French Minister Delegate of Artificial Intelligence and Digital Technologues, Mrs Clara Chappaz at ATx 2025.
Singapore and US spearhead the Global Cross-Border Privacy Rules (CBPR) Certification for businesses to facilitate trusted cross border data flows, providing access to 9 economies with about 40 trillion in market size. This was developed by the Global CBPR Forum, of which Singapore serves as deputy chair. Companies can apply for the new Certification from 2 June 2025, which will allow organisations to demonstrate compliance with internationally recognised data protection standards.
Singapore's Enigma Health partners with Roche and ST Engineering
Enigma Health, a spinoff from SingHealth Duke-NUS Academic Medical Centre, signed two strategic partnerships on 27 May 2025, with Roche and ST Engineering's Enterprise Digital to expand its capabilities. This was announced by Mdm Rahayu Mahzam, Minister of State for Digital Development and Information at the "Scaling and Sustaining Healthcare with Gen AI" symposium held at ATxSG. These two new partnerships will accelerate clinical trials, market access, insights and business intelligence to help with patients' care and access to novel drugs and digital technologies.
Enigma Health's flagship product, Enigma, is Singapore's home-grown sovereign healthcare agentic AI platform developed by a team of clinicians and AI scientists to improve workflow optimisation, streamlining data-intensive and time-consuming processes while ensuring robust data security and regulatory compliance. It has been piloted at SingHealth institutions.
IMDA and Thailand's National Broadcasting and Telecommunications Commission (NBTC) have signed a Memorandum of Understanding (MoU) to combat telecommunications fraud across ASEAN.
The MOU was exchanged between Ms Jasmin Lau, Minister of State at the Ministry of Digital Development and Information and Dr Sarana Boonbaichaiyapruck, Chairman of the NBTC at the sidelines of ATxSG 2025.