NUS Launches AI Sense Maker Chatbot to Unlock 150,000 Rare Collections and Research Papers

Singapore's oldest university is betting that AI can solve one of academia's most persistent problems: finding the right source material in a sea of digitised knowledge. The National University of Singapore has launched AI Sense Maker, an AI chatbot powered by OpenAI's GPT 4.1 and retrieval-augmented generation, that makes 150,000 rare collections and internal research papers searchable through natural language queries — cutting what typically takes researchers two weeks down to about 30 minutes.

The platform, launching in August for NUS students, faculty, and alumni, draws on a vast repository that includes the university's internal research papers in engineering, life sciences, medicine, humanities, and computing, alongside rare collections dating back to the 17th century. Among the digitised treasures are Lat Pau, the first Chinese-language daily newspaper published in the Nanyang region (launched in 1881), the first printed edition of Sir Stamford Raffles' The History of Java (published 1817), and the original musical compositions and scripts of beloved entertainer Wang Sha. Each answer produced by the chatbot includes cited sources, topic summaries, related concepts, and a reference list.

What distinguishes AI Sense Maker from general-purpose chatbots is its guardrails. NUS University Librarian and Associate Professor Natalie Pang noted that popular AI tools often pull from unverified internet sources — a problem in higher education where source credibility is paramount. AI Sense Maker is restricted to its curated database and cannot search the public internet, while the RAG architecture prevents the model from fabricating answers that don't exist in the source material. The tool also features a knowledge graph that maps conceptual connections across disciplines, helping researchers identify thematic gaps in their inquiry.

The launch reflects a broader push by Singapore's universities to develop institution-specific AI tools that serve their unique research ecosystems. NUS chief IT officer Tan Shui-Min highlighted how the knowledge graph feature could boost early-stage research by making themes and gaps visible that might otherwise take weeks of literature review to identify. NUS is already planning upgrades including extended conversation memory for follow-up queries and voice-to-text transcription, suggesting this is the first iteration of what could become a more comprehensive AI research assistant.

Why it matters for Singapore: AI Sense Maker is a concrete example of how Singapore's research institutions are building AI tools purpose-fit for their own needs — not just adopting off-the-shelf solutions. For a country that has positioned itself as an AI hub, the quality of its research infrastructure matters, and tools like this strengthen the value proposition for academics and researchers choosing Singapore. The focus on source credibility and cited answers also sets a useful precedent for how AI should be deployed in knowledge discovery, particularly in education settings where hallucination risks have slowed adoption.