Get Started
Home
Topics
Search
Library
Research questionHow can affective AI be evaluated using long-term, naturalistic, passively sensed workplace data?Most affective-computing systems are evaluated on short, controlled datasets, making it difficult to separate person-specific, team-level, and seasonal variation in real workplaces. Conventional classification metrics can also miss demographic bias and poorly represent outcomes such as employee turnover.
AI
Computer Vision
Evaluation & Benchmarks
Image & Video Processing
Machine Learning
Research Paper
Latest papersRecent research connected to this question, newest first.WELD: The First Naturalistic Long-Period Small-Team Workplace Emotion Dataset for Ubiquitous Affective ComputingWELD provides 733,780 per-frame seven-class facial-expression probability vectors from 49 employees at a Chinese software company over 30.1 months, collected through a fully passive protocol within stable small teams. It supports within-person longitudinal analysis, within-team relational analysis, fairness auditing of facial-expression recognition, and comparison of turnover prediction with survival-aware metrics. The dataset uses four access tiers; only aggregated probabilities are publicly downloadable. Evidence comes from this single workplace and employee cohort.research paper · Sep 3, 2026
Related questions
How can software organizations adopt AI without imposing psychological and professional costs on engineers?How can we measure an AI agent’s tacit alignment with a human when objectives, communication, and feedback are limited?Can deterministic artificial affective processing produce hedonic place preference without conscious feelings?How can AI-derived respondent measures recover within- and between-group effects when conventional surveys are unavailable?