Building AI for Our Communities: Useful, Accessible, and Safe

​As AI continues to advance rapidly, the key question is no longer just what AI can do—but how we ensure that the AI applications we build meaningfully serve our communities.

​This session brings together practitioners, researchers, and ecosystem partners to explore what it means to design and deploy AI systems that are useful, accessible, and safe in real-world contexts. Beyond leveraging open-source resources such as SEA-LION models, we will also explore how communities can contribute to open-source efforts to ensure AI better reflects local needs, languages, and cultural contexts, and how we can bridge the gap between innovation and societal impact.

​Programme Agenda

​11.30am: Registration, Lunch & Networking

​12.55pm: Opening Remarks

​1.05pm: Introduction to SEA-LION and New Model Capabilities

​1.25pm: Project Aquarium New Release

​1.55pm: [Panel Discussion] Local Data, Global Models: Rebalancing the AI ecosystem

Panel details: AI development has generally relied on a "top-down" approach, often overlooking the linguistic and cultural nuances of diverse regions. This panel challenges the status quo, asking: How do we build competitive AI while remaining respectful and responsible to the communities providing the data? Our speakers will share practical frameworks for data collection that prioritize inclusivity and mutual benefit.

​Featuring perspectives from global AI experts, the session will also explore strategies for building high-quality models in resource-constrained environments. Join us to discover how community-led initiatives are leveling the playing field and why open, community-driven data is essential to the future of AI.

​2.55pm: Ecosystem Showcase

​3.25pm: Closing Remarks

​3.35pm: Networking

​Don’t miss the chance to experience SEA-LION in action at our interactive AI zone!