Research map / Large language models / AI and society
Role of Human-AI Interaction in Selective Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Show deferral status while withholding the AI prediction | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images | human accuracy: 58.4% -> 61.9% | Helped |
| Show deferral status, averaged across prediction-message conditions | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images | human accuracy: 57.4% -> 60.4% | Helped |
| Show the AI prediction, averaged across deferral-status conditions | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images | human accuracy: 60.2% -> 57.8% | Hurt |
| Show the incorrect AI prediction without showing deferral status | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images where AI is incorrect | human accuracy: 50.6% -> 41.9% | Hurt |
| Show AI predictions rather than no message | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images | conformity score: 0 -> 0.08 | Helped |
| Provide AI predictions on images the model classifies correctly | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images where AI is correct | human accuracy: +5.1% | Helped |
| Provide AI predictions on images the model classifies incorrectly | Human reviewers in the selective-prediction workflow · Snapshot Serengeti deferred images where AI is incorrect | human accuracy: -8.7% | Hurt |
| Rely on historical human-label performance without modeling the communication message | Selective-prediction system evaluation · Snapshot Serengeti deferred images | human accuracy | Mixed |
Closest papers by meaning
- Utilizing Human Behavior Modeling to Manipulate Explanations in AI-Assisted Decision Making: The Good, the Bad, and the Scary (NeurIPS 2024 · similarity 0.88)
- Updates in Human-AI Teams: Understanding and Addressing the Performance/Compatibility Tradeoff (AAAI 2019 · similarity 0.88)
- Human Expertise in Algorithmic Prediction (NeurIPS 2024 · similarity 0.87)
- Decoding AI’s Nudge: A Unified Framework to Predict Human Behavior in AI-Assisted Decision Making (AAAI 2024 · similarity 0.87)
- Mecha-nudges for Machines (NeurIPS 2026 · similarity 0.86)
- Large Language Models Assume People are More Rational than We Really are (ICLR 2025 · similarity 0.86)
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.