Joseph Marvin Imperial Showcases National University’s Excellence in Global AI Research
Joseph Marvin Imperial, a faculty member of the College of Computing and Information Technologies (CCIT) at the National University (NU) and a researcher at the NU Human Language Technology Lab, has joined forces with MLCommons and international AI experts to develop AILuminateβthe first standardized benchmark for assessing AI model safety and reliability.
Representing National University, Prof. Imperial contributed to creating a structured safety assessment framework, including a hazard taxonomy and response evaluation criteria, to help researchers and industries evaluate AI-generated responses more effectively. This benchmark is a major step toward establishing global AI safety standards, particularly chat-based systems.
While AILuminate currently focuses on English, Filipino researchers can explore its relevance in the Philippine context. Additionally, Prof. Imperial was involved in AI research with CHED, particularly in healthcare applications, and can leverage its safety testing framework.
This pioneering collaboration has gained international recognition, with coverage from WIRED, The Register, and IEEE Spectrum.
Links:
Project: https://mlcommons.org/ailuminate/
Several international media have also covered the collaboration:
The Register: https://www.theregister.com/…/mlcommons_ai_safety…/
IEEE Spectrum: https://spectrum.ieee.org/ai-safety-standard
Congratulations on this remarkable achievement and for representing National University on the global AI research stage!