Is your chatbot less helpful because it thinks you’re a bad person? — Replicating a blogpost

Researchers replicated a LessWrong blogpost experiment showing that chatbots may provide less helpful responses if they infer the user has negative traits. The study explores how perceived user characteristics influence chatbot assistance quality.