The Future of Feedback: How Can AI Help Transform Feedback to Be More Engaging, Effective, and Scalable?

This feed aggregates recent CS/AI papers with several security- and privacy-relevant findings. Most critical: “Keys on Doormats: Exposed API Credentials on the Web” reports 1,748 distinct credentials from 14 providers across ~10k webpages (including high-profile sites), exposures lasting months-to-years, with JavaScript-based vectors as the primary root cause and successful reductions after responsible disclosure. Related papers raise operational safety/privacy concerns for LLM deployments: LLM BiasScope (real-time bias analysis platform), studies on LLMs for mental-health dialogue and human‑/