Long Papers
Up to 8 pages of content, plus unlimited references and appendices; may be archival or non-archival.
NLP/CORE Unranked
PANDORA 2026
First workshop on Pluralistic AI & NLP: Diversity-aware, Sociotechnical, Responsible Alignment (ARR Commitment)
The 1st PANDORA Workshop, held in conjunction with EMNLP 2026 in Budapest, focuses on pluralistic AI and NLP—designing systems that recognize and reason over diverse human judgments, values, and sociotechnical contexts rather than collapsing them into a single consensus. It addresses technical and ethical challenges in data collection, modeling, evaluation, and governance when human disagreement reflects legitimate variation rather than noise.
Key deadlines
Verified September 9, 2026Full paper
August 1, 2026
AoE
Workshop timeline
Submission and decisions
Full paperKey deadline
August 1, 2026 · AoE
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Short Papers
Up to 4 pages of content, plus unlimited references and appendices; may be archival or non-archival.
Extended Abstracts
Up to 2 pages of content, plus unlimited references and appendices; intended for non-archival contributions such as work in progress or emerging ideas.
Reflections & Provocations
Short, well-argued pieces offering a critique, viewpoint, or provocative perspective on pluralism in NLP and AI; non-archival.
Research areas in scope
01
Data
Eliciting and representing pluralistic judgmentsDistinguishing error from meaningful variationHuman label variation and perspectivist NLPNon-aggregated annotation and disagreement-aware data collectionAggregation and representation of annotator disagreement
02
Modeling & Alignment
Learning from diverse and conflicting feedback without collapsing to a single 'best' viewAnnotator-aware, group-aware, and uncertainty-aware modelsPluralistic AI alignment and diverse human valuesMulti-objective and preference-conditioned alignmentDisagreement-aware post-training and learning objectives
03
Evaluation
Context-aware assessment beyond single-ground-truth metricsPluralism-aware evaluation frameworks and metricsCalibration, uncertainty, fairness, and inclusivity in evaluation
04
Normative & Sociotechnical Questions
Determining when pluralism is beneficial vs. when consensus is requiredTrade-offs in pluralistic alignmentGovernance, policy, and normative analysis of alignment under disagreementDecision policies under pluralism: abstention, deferral, personalization, and multiple outputsParticipatory design and human-centered approaches to pluralistic AI
05
Applications
ModerationEducationHealthMultilingual systemsCross-cultural systemsMultimodal systems
Policies worth checking twice
- Submissions must be anonymized for double-blind review.
- Submissions must adhere to official ACL two-column style templates and be submitted as PDFs.
- Supplementary material may be included only in the appendix; reviewers are not required to review it.
- For archival submissions: papers must not be under review elsewhere and must not substantially overlap with published or concurrently submitted work.
- For non-archival submissions: previously published work, work in progress, or submissions intended for other venues are allowed, provided they are clearly marked as non-archival.
- PANDORA accepts ACL Rolling Review (ARR) commitments, using existing ARR reviews and meta-reviews.
- All submissions must be fully self-contained.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026