Culturally (Un)Aware AI
Abstract
It is widely known that AI defaults to Western values and cultures, underrepresenting other cultures. In this talk, I will first show the downstream impacts of this cultural underrepresentation. I will discuss how current explainable AI techniques fail to include the cultural context of community health workers (CHWs) in India. Then, I will show how, conversely, carefully designed AI can help make technology more culturally appropriate. I will do so by discussing the design of a culturally appropriate conversational agent to facilitate the deliberation of harmful content on WhatsApp in the Global South.
Resource for Attendee: None
Seminar Details
Presenter Bio
Dhruv Agarwal is a second-year Ph.D. student at Cornell University advised by Aditya Vashistha. He is a mixed-methods researcher working on Culturally Aware AI and has looked at this problem from various lenses ranging from explainable AI to online harmful content. He has won paper awards at CSCW and COMPASS and is an inaugural Quad Fellow. Born and brought up in India, he graduated magna cum laude in Computer Science from Ashoka University, India, and then spent two years at Microsoft Research India working on air pollution. He is passionate about the real-world impact of research and hopes to bridge the technological disparity in society. Outside work, Dhruv is an avid sportsperson and has played tennis, squash, and soccer competitively.