ABOUT THIS ISSUE

How was this newsletter synthesized?

Methodology

This newsletter is generated by an AI pipeline (leveraging Anthropic Sonnet 4.5 & Haiku 4.5) that processes the metadata and abstracts of every new arXiv HCI paper from the past week—100 this issue. Each paper is scored on three dimensions: Practice (applicability for practitioners), Research (scientific contribution), and Strategy (industry implications), with scores from 1-5. Papers passing threshold are grouped into topic clusters, and each cluster is summarized to capture what that body of research is exploring.

Selection Criteria

The pipeline builds a curated selection that balances high scores with topic diversity—and deliberately includes at least one 'contrarian' paper that challenges prevailing assumptions. This selection is then analyzed to identify key findings (patterns across multiple papers) and surprises (results that contradict conventional wisdom). A narrative synthesis ties the week's research together under a unifying frame.

Key Themes Discovered

Field Report: ai-interaction

Measuring Trust, Judgment, and Drift

Does AI perform worse, or do people judge it more strictly? This batch asks how users calibrate reliance on LLM advice, whether evaluation standards drift, and how agentic systems set their own autonomy. Papers pair controlled user studies and logged-vote analyses with interviews, mechanistic probes, and small benchmarks. Most work treats human judgment as the unit under test, not the model alone. Primarily relevant for HCI researchers studying trust, leaderboard designers, and teams deploying advisory or agentic assistants.

1/10