KORIKA and AI Singapore Launch Project NUSA for Trusted, Locally Grounded AI in Indonesia

Indonesia's AI research collaboration body KORIKA and AI Singapore (AISG) have launched Project NUSA, short for Network of Unified and Secure AI, a joint effort to give Indonesia trustworthy AI grounded in local language, rules and context, Suara.com reported on October 22. The project was unveiled on October 21 at the AI Impact Summit in Sanur, Denpasar, Bali, held during Bali Tech Week.

Coordinating Minister for Economic Affairs Airlangga Hartarto backed the initiative in virtual remarks, saying broad use of AI requires trust and that Project NUSA aims to build that foundation through strong governance and relevant knowledge. The partners present it as a way to raise industrial productivity, attract investment and open new sources of income.

Project NUSA rests on two pillars. The first, open-source AI governance, covers safe and responsible use, including access controls, oversight of AI agent activity and accountability. The second, open data and AI-ready knowledge, turns data that can legally be shared into trusted knowledge so AI systems better understand Indonesian language, regulation and local context. KORIKA chair Hammam Riza said the body will orchestrate Indonesia's AI ecosystem while drawing on AI Singapore's experience. He argued that enthusiasm for generative AI content has overlooked governance and data originality, and that much of today's generative AI data lacks authentic Indonesian value.

What it means in practice

For Indonesian developers, the interesting part is the data pillar. Models trained mostly on foreign text struggle with local regulations and languages, which is why projects such as GoTo's open-source LLM for Indonesian and regional languages drew attention. A shared, legally cleared knowledge base could help smaller teams that cannot assemble such data alone. William Tjhi, deputy director of AI products at AI Singapore, said success depends not only on models but on knowledge access, orchestration and governance, which he said grows harder as AI moves faster.

The launch also fits Jakarta's wider rule-making. Komdigi has been working on a national AI coordination task force with three risk tiers in draft presidential rules and on a ministerial rule requiring labels on synthetic content, and oversight of AI agents is likely to be a recurring theme in both.

What to watch: which datasets Project NUSA releases first, which companies and universities sign on, and whether its governance tools are adopted by government agencies deploying AI agents.