Ziad Al-Halah develops efficient multimodal AI assistants that learn with constrained resources and understand text, images, and video. His research addresses key environmental and accessibility challenges in modern AI by reducing supervision costs, lowering power consumption, and creating practical tools for education. For example, his AI assistants can separate audio-visual sources to generate clearer captions for hearing-impaired students or analyze classroom recordings to help users quickly locate specific moments. “I want to enrich students’ and instructors’ experiences and help shape AI as a force for the public good,” Al-Halah said. He also plans to expand this work through a new course on multimodal AI. “By combining cutting-edge research with ethical considerations, my goal is to prepare students not only to innovate in AI, but also to apply it responsibly.”
