New evaluation methods and benchmarks
Methods and benchmarks for evaluating human-agent systems in deployed settings.
Machine Learning/CORE Unranked
NeurIPS-26 HAIC Workshop 2026
2026 NeurIPS Workshop Human-AI Coevolution
The HAIC @ NeurIPS 2026 workshop focuses on developing rigorous empirical evaluation methods for human-agent teams in the agentic era, particularly in high-stakes domains like healthcare, mental health, aviation, and finance. It builds on the inaugural ICLR 2025 workshop by narrowing its scope to three interrelated themes centered on validity, expert disagreement, and adaptive evaluation as systems and users coevolve.
Key deadlines
Verified September 9, 2026Full paper
August 30, 2026
AoE
Workshop timeline
Submission and decisions
Full paperKey deadline
August 30, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Datasets
Datasets capturing human-agent interaction in deployed or longitudinal settings.
Case studies
Case studies from high-stakes domains such as healthcare, mental health, aviation, and finance.
Reproducible critiques and replications
Reproducible critiques or replications of existing agentic benchmarks.
Position papers
Position papers advancing validity-centered or coevolutionary evaluation frameworks.
Research areas in scope
01
Themes
Validity of evaluation in deployed contextsExpert disagreement and the limits of human feedbackAdaptive testing and continual evaluation for coevolving systems
Policies worth checking twice
- Submissions are double-blind via OpenReview.
- Main text must be at most 9 pages; references and appendices do not count toward the limit.
- The workshop is non-archival: accepted papers may be submitted to conferences or journals afterwards.
- Posting to preprint servers like arXiv is permitted before and during review.
- Work that has appeared on non-peer-reviewed websites or presented without proceedings may be submitted.
- Large language models may be used as assistive tools, but cannot be authors; authors are responsible for all content, including model-generated material.
- Submissions may be withdrawn before notification.
- Reviewers are subject to NeurIPS conflict-of-interest policies and cannot review papers from their own institutions.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026