Research map / Large language models / AI and society
Fair Decision Utility in Human-AI Collaboration: Interpretable Confidence Adjustment for Humans with Cognitive Disparities
Jiashi Gao, Kexin Liu, Xinwei Guo, Junlei Zhou, Jiaxin Zhang, Xiangyu Zhao, Guanhua Chen, Xin Yao, Xuetao Wei
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Art task | EIAE: 0.0345 -> 0.0209 | Improved |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Census task | MIAE: 0.5275 -> 0.1062 | Improved |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Sarcasm task | MIAE: 0.5702 -> 0.0458 | Improved |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Cities task | EAE: 0.0003 -> 0.0007 | Hurt |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Art task | utility disparity (t-test): t = 9.486, p = 0.000 | Improved |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Cities task | utility (t-test): t = 0.345, p = 0.731 | No clear effect |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Sarcasm task | utility (t-test): t = -12.187, p = 0.000 | Hurt |
| use cognition-aware instead of cognition-unaware multicalibration | AI confidence adjustment · Census task | utility disparity (t-test): t = 27.556, p = 0.000 | Improved |
| add cognition-unaware multicalibration to unadjusted confidence | AI confidence adjustment · Art task | utility disparity | Hurt |
| adjust AI confidence by multicalibration (either variant) versus no adjustment | AI-assisted decisions · four tasks | decision utility | Improved |
| use cognition-aware multicalibration under higher monotonicity violation (sampled noisy set) (ablation) | AI confidence adjustment · Census task (noisy set) | utility disparity (t-test): t = 1.010, p = 0.326 | No clear effect |
| use cognition-aware multicalibration under higher monotonicity violation (sampled noisy set) (ablation) | AI confidence adjustment · Sarcasm task (noisy set) | utility disparity (t-test): t = 55.487, p = 0.000 | Improved |
Closest papers by meaning
- Learning to Decide with AI Assistance under Human-Alignment (NeurIPS 2026 · similarity 0.91)
- Measuring Error Alignment for Decision-Making Systems (AAAI 2025 · similarity 0.89)
- Position: Towards Bidirectional Human-AI Alignment (NeurIPS 2025 · similarity 0.88)
- Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties (AAAI 2024 · similarity 0.88)
- Trustworthy AI Must Account for Interactions (NeurIPS 2026 · similarity 0.87)
- AI Alignment with Changing and Influenceable Reward Functions (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
