Denny George, Tattle

Artificial Intelligence is having a cultural moment. Emergent behaviour of Large Language Models has perplexed everyone about the good and bad disruptions they can cause. These models reduce the time and effort needed to create realistic looking media in text or image form. This has implications for anyone interested in the information disorder space. This session will do a survey of what’s been made possible by the state of the art in Machine Learning. We will focus on the specific challenges of content production and harm mitigation with these models in the South Asian context. We will share experiments from our work and other tools and community efforts around mitigating harms.