AISG Partners with US National Science Foundation on Trustworthy AI Research

AI Singapore has launched a collaborative research project with the US National Science Foundation, bringing together researchers from Nanyang Technological University, Singapore Management University, and the Massachusetts Institute of Technology to tackle one of AI's hardest problems: making human-AI collaboration systems truly trustworthy. The project aims to build AI systems that can handle complex decision-making through natural language while remaining transparent, explainable, and reliable.

The research targets a fundamental limitation in current AI collaboration tools — their inability to handle combinatorial optimisation problems through natural conversation. Led by Professor Jie Zhang of NTU and Assistant Professor Zhiguang Cao of SMU on the Singapore side, and MIT's Associate Professors Jacob Andreas and Cathy Wu on the US side, the team is developing a framework that can translate imprecise human language into formal optimisation problems, solve them reliably, and explain the reasoning back in plain language.

The project has three core goals: formulating complex optimisation problems from ambiguous human instructions, ensuring robust AI-assisted problem-solving across diverse conditions, and returning explainable feedback that builds user trust. Potential applications span logistics, manufacturing, finance, and personal planning — from intelligent calendar assistants to supply chain optimisation tools that can respond to natural language requests.

Why it matters for Singapore: This AISG-NSF collaboration places Singapore's top research universities at the frontier of trustworthy AI research, alongside an MIT team that sets global benchmarks in human-AI interaction. For Singapore's Smart Nation ambitions, research that makes AI systems more trustworthy and interpretable directly supports broader enterprise and government adoption — addressing one of the biggest barriers to AI deployment at scale in the city-state.