Our Research

At Fire AI Lab , we push the boundaries of artificial intelligence to solve real-world problems and enhance digital safety. Our research focuses on two main areas: Applied AI , where we leverage large-scale data to drive breakthroughs in various fields, and Safe AI , where we develop cutting-edge techniques to detect and mitigate emerging AI-driven threats. Our work is rooted in open science, ensuring our findings, tools, and datasets are accessible to researchers and practitioners worldwide. Applied AI

Applied AI

We harness AI and big data to generate insights that make a tangible impact across disciplines. Our projects include:

  • Measuring media bias by analyzing millions of headlines with LLMs
  • Deciphering historical manuscripts using AI
  • Identifying abnormal vertices in networks, whether they represent malicious users in online systems or abnormal proteins in biological data
  • Enhancing government transparency by analyzing parliamentary proceedings
  • Improving urban planning through AI-driven insights into transportation and infrastructure
Safe AI

As AI-generated content and security threats evolve, we are at the forefront of ensuring digital safety. Our research in Safe AI includes the following:

  • Identifying fake images and videos to combat misinformation
  • Detecting malicious code to enhance cybersecurity
  • Developing AI-driven fraud detection to uncover fake profiles and online deception
  • Enhancing LLM security by developing tools to test for jailbreak vulnerabilities, ensuring AI models remain safe

Open Code and Data

We believe that impactful research should be transparent, reproducible, and accessible. That’s why we openly share many datasets, algorithms, and publications—empowering researchers and practitioners to build on our work and drive further innovation.

Learn more

Explore our Projects and Datasets to dive deeper into our research and access the tools we’ve developed.

logo