HEAL@CHI'26 2026

Deadlines
NLP/CORE Unranked

HEAL@CHI'26 2026

Third Workshop on Human-Centered Evaluation and Auditing of Language Models: AI Agents-in-the-Loop

Apr 13 2026Barcelona, SpainOfficial workshop site Site reachable

HEAL@CHI'26 is a workshop returning for its third iteration at CHI 2026, focused on addressing the evaluation crisis in large language models through human-centered approaches. It brings together HCI and AI researchers to develop methods, tools, and frameworks for auditing LLMs, with a special theme on AI agents-in-the-loop and how to balance automation with human agency.

Official CFP Back to deadlines Verified September 9, 2026

Key deadlines

Verified September 9, 2026

Full paper

February 20, 2026

AoE

Workshop timeline

Submission and decisions

Full paperKey deadline

February 20, 2026 · AoE

Paper fit

Contribution paths

A strong submission should clearly identify its contribution and evaluate it appropriately.

Encore of published work (non-archival)

Non-archival submissions that do not require anonymization and undergo a jury review process

Research areas in scope

01

General Topics

Empirical understanding of stakeholders' needs and goals in LLM evaluation and auditingHuman-centered evaluation and auditing methods for LLMsTools, processes, and guidelines for LLM evaluation and auditingDiscussion of regulatory measures and public policies for LLM auditingEthics in LLM evaluation and auditing
02

Special Theme: AI Agents-in-the-Loop

Task allocation and workflow integration between human evaluators and AI agentsImpact of hybrid human-AI approaches on vulnerability discoveryMeta-evaluation frameworks for assessing trustworthiness of AI agents as auditing toolsMethodologies for evaluating complex agent behaviors (multi-step reasoning, tool use, emergent patterns)Design patterns and safeguards for maintaining human oversight in automated evaluationEmpirical studies of agent-assisted evaluation in practice

Policies worth checking twice

  • Submissions must be 2-6 pages in ACM double-column format, excluding references
  • Review process is double-blind, except for encore submissions
  • Encore submissions do not need to be anonymized and undergo a jury review
  • Authors must use \documentclass[sigconf,anonymous]{acmart} for submissions
  • New OpenReview profiles without institutional email may take up to two weeks to be moderated
  • Authors are encouraged to help with the review process

Official sources

Compiled from the official call for papers. The organizers’ pages remain authoritative.

Last verified September 9, 2026