Stanford Daily: Stanford AI researchers develop new algorithm for more inclusive content moderation

Stanford Daily: Stanford AI researchers develop new algorithm for more inclusive content moderation. “Researchers at the Stanford Institute for Human-Centered Artificial Intelligence (HAI) proposed a jury learning algorithm that aims to improve online content moderation last spring, with hopes to deploy their system on platforms like Reddit and Discord in the near future.”

Leave a Reply

%d bloggers like this: