HK AI Scene

The Research You're Missing: AI Papers from HKUST, HKU, and CUHK

Hong Kong AI Podcast/2026-03-07/6 min read/HKUSTHKUCUHKResearchAI PapersHong Kong

Hong Kong's three leading universities — HKUST, HKU, and CUHK — are producing AI research that competes with anywhere in the world. But unless you're reading NeurIPS proceedings or following specific lab groups, you probably don't know about it.

Here's what's coming out of Hong Kong's AI labs.

HKUST: #17 Worldwide and Climbing

HKUST ranks 17th globally in "Data Science and AI" according to QS 2025 — the highest in Hong Kong. Their AI research spans several groups, but two projects stand out.

Aivilization: 100,000 AI Agents in a Virtual Society

This is HKUST's headline project and it's genuinely wild. Aivilization is the world's largest AI multi-agent social simulation — 100,000 AI agents interacting in a virtual society, forming relationships, making decisions, and creating emergent social structures.

The research goal is understanding how AI agents behave at societal scale. How do they cooperate? How do conflicts emerge? What social structures form? The applications range from policy simulation to game design to understanding the future of multi-agent AI systems.

It's also an educational platform — students can observe and interact with the simulation to learn about AI and social dynamics. HKUST has positioned it as both a research tool and an AI literacy initiative.

Leading HKGAI

HKUST leads the Hong Kong Generative AI R&D Center, coordinating five universities and producing HKGAI V1, HKChat, and the suite of specialized applications (LexiHK, HKPilot, HKMeeting, etc.). This is the most concrete example of university research translating directly into public-facing AI products in Hong Kong.

HKU: Autonomous AI Scientists

AI-Researcher: NeurIPS 2025 Spotlight

HKU's Data Intelligence Lab developed "AI-Researcher" — a system designed for autonomous scientific innovation. It can formulate research questions, design experiments, analyze results, and write papers. It was selected as a NeurIPS 2025 Spotlight paper, which is a significant recognition.

The system doesn't replace researchers, but it accelerates the research process. Think of it as an AI research assistant that can run through hundreds of hypothesis variations while the human researcher focuses on higher-level direction.

The code is open source on GitHub (HKUDS/AI-Researcher), which means other labs — including those in Hong Kong — can build on it.

Medical AI

HKU's medical AI group, led by Prof. Lequan Yu, has published in Nature Communications, Cell Genomics, ICLR, NeurIPS, and AAAI across 2025-2026. Their work spans medical imaging, genomics, and clinical decision support.

Hong Kong's position as a medical hub — with world-class hospitals and a mix of Eastern and Western medical traditions — gives medical AI researchers here unique datasets and clinical partnerships.

CUHK: Trustworthy and Embodied AI

The Interdisciplinary AI Research Institute

CUHK established this institute in partnership with the Shanghai AI Laboratory. It's a significant collaboration that bridges Hong Kong's research strengths with mainland China's AI ecosystem.

Focus Areas

CUHK-Shenzhen's School of AI has three research pillars that feel particularly forward-looking:

Trustworthy AI — As AI gets deployed in critical systems, ensuring it's reliable, fair, and safe becomes essential. This research is directly relevant to Hong Kong's regulated industries: finance, healthcare, legal.

Embodied AI — AI that interacts with the physical world through robotics and spatial understanding. Given Hong Kong's density and complex urban environment, this has natural local applications.

AI for Science — Using AI to accelerate scientific discovery across disciplines. This complements HKU's AI-Researcher work.

What Connects All of This

Three patterns emerge from Hong Kong's AI research landscape:

1. Application-oriented. HK's AI research tends to target practical outcomes — HKChat for citizens, medical AI for hospitals, legal AI for lawyers. There's less "research for research's sake" and more "research that ships."

2. Collaborative. HKGAI involves five universities. CUHK partners with Shanghai AI Lab. The research ecosystem is highly interconnected, which is natural in a small city where every researcher is one degree of separation from every other.

3. Leveraging open source. HKGAI V1 is built on DeepSeek. AI-Researcher is open-sourced on GitHub. The HK research community has fully embraced the open-source AI paradigm, which makes sense when your city is blocked from US proprietary tools.

Why You Should Care

If you're building AI products in Hong Kong, these university labs are your local talent pipeline, your potential research partners, and your source of ideas. The researchers publishing at NeurIPS and Nature are training the PhD students who'll be your hires or co-founders in a few years.

And if you're outside Hong Kong, the research coming out of these labs is a signal: this city is producing world-class AI work, and it's only accelerating.

Sources

Want to hear directly from the researchers building Hong Kong's AI future? Subscribe to the Hong Kong AI Podcast.

Stay in the loop

Get notified when we publish new articles and episodes. No spam, just signal.

Something out of date or wrong? AI moves fast and we want to get it right. Let us know at contact@hongkongaipodcast.com