aphotostory - stock.adobe.com
Sea-Lion powering AI tools for migrant workers, local businesses
The inaugural Sea-Lion summit showcased real-world applications of Southeast Asia’s first homegrown large language model, which now supports multimodal inputs and can run on a standard laptop
From a smartphone platform helping migrant workers in Thailand file grievances in their native tongues to helping businesses navigate the nuances of local dialects, Southeast Asia’s first artificial intelligence (AI) model is moving from the lab to the real world.
These applications were showcased today at the inaugural Sea-Lion Summit organised by AI Singapore (AISG), highlighting how the open-source model is closing the gap left by Western-centric AI models that often struggle with the region’s diverse languages and cultural contexts.
Since it debuted in November 2023, Sea-Lion, an acronym for Southeast Asian Languages in One Network, has been gaining traction across the region. To date, it has shared over one trillion Southeast Asian language tokens across 13 regional languages and local dialects.
AISG recently teamed up with Alibaba Cloud on Qwen-Sea-Lion-v4 – a version of the model built on the Chinese hyperscaler’s Qwen3-32B base model – that can run on a consumer-grade laptop with 32GB of RAM. This is a critical feature for the region, where many small and medium-sized enterprises and developers lack access to industrial-grade graphics processing unit clusters.
To ensure Sea-Lion, which now supports multimodal inputs, continues to reflect the region’s voices, cultures, and needs, AISG has also developed the Aquarium open data platform together with Google that allows users, including researchers, developers, policymakers and grassroots organisations to contribute, browse, and collaborate on language datasets across the region.
Solving real-world problems
The summit highlighted how organisations are already deploying Sea-Lion to address social issues. For example, UN Women is using the model for its Talking Justice initiative, which aims to support migrant workers in Thailand by providing a multilingual, AI-supported grievance platform.
“Workers will be enabled to file complaints safely, privately, and in their own language, directly from their phones or community centres,” said Cate Sumner, senior consultant for digital public goods for gender justice at UN Women.
Sumner added that the system integrates automated translation and conversational guidance to link workers to legal or psychosocial services, removing “critical friction points” in the current system.
On the commercial side, companies like NCS and Good Bards are using the model to build scalable, multilingual AI platforms and services for their clients.
Alan Ho, CEO and co-founder of Good Bards, said integrating Sea-Lion into the company’s marketing operating system gives clients a distinct advantage in engaging the Southeast Asian region through a more localised lens.
“The importance of local context in customer engagement cannot be overstated, and with Good Bards, customers can easily leverage Sea-Lion’s capabilities to engage their audience effectively,” he added.
To encourage local innovation, AISG also organised the Pan-Southeast Asia AI Developer Challenge, which concluded at the summit. Supported by tech giants including Amazon Web Services, Google Cloud, and Nvidia, the challenge attracted nearly 800 participants across 235 teams.
Winning solutions included tools to improve financial literacy for unbanked communities and multi-dialect health advisories that bridge linguistic gaps across the region.
Leslie Teo, senior director of AI products at AISG, said the innovation shown at the challenge proves what an engaged community can achieve when empowered by an open, regionally focused foundation.
“We hope this summit inspires developers, researchers, and enterprises to build and deploy the next generation of impactful AI solutions that addresses the many challenges and opportunities in Southeast Asia,” he said.
Read more about AI in ASEAN
- Malaysia’s Ryt Bank is using its own LLM and agentic AI framework to allow customers to perform banking transactions in natural language, replacing traditional menus and buttons.
- By unifying data on the Databricks platform, Vietnam’s Techcombank has built AI capabilities to deliver hyper-personalised offers to 15 million customers and expand its footprint beyond its traditional affluent base.
- Researchers at the National University of Singapore have created a wearable device that combines a camera with conversational AI powered by Meta’s Llama models to give sight to the visually impaired.
- Indonesia’s GoTo has migrated half its infrastructure to Alibaba Cloud, paving the way for AI initiatives to solve real-world business problems and support local languages.
