1/ 🚀 We’re excited to launch ECHO (Evaluation of Chat, Human Behavior, and Outcomes), an open, low-code research platform for reproducible user studies on conversational AI, search, and information access. Built to lower technical barriers, ECHO enables rigorous human-centered
Latest posts 5
(1/2) Our new survey work systematically reviews user studies on human-centered explainability—AI explanations designed for and assessed by real people, not just algorithms. Our findings show how it’s defined, built into interfaces, and measured, revealing what truly helps users
🧐 Need IR user-interaction data but stuck on sketchy docs & privacy hurdles? Our just-accepted JASIST paper shows how #IIR scholars actually find, judge & reuse datasets, pinpoints key reuse drivers and challenges, and offers metadata + privacy frameworks to unlock those silos.
1/4 🚀 New Research on LLM Personality and Cognitive Bias 🤖 Our latest study with @jiangenhe systematically investigates how personality traits influence cognitive biases in LLMs and assesses the effectiveness of bias mitigation strategies across different architectures.