Workshop Overview
Recent advances in Generative AI (GenAI) and Large Language Models (LLMs) have transformed information access, knowledge discovery, decision support, and human-computer interaction. Despite remarkable progress, existing evaluation frameworks and benchmarks often fail to adequately assess how these systems perform across diverse cultural, linguistic, socioeconomic, and community contexts. As GenAI technologies become increasingly integrated into education, healthcare, government, social services, and everyday life, ensuring cultural competence, fairness, inclusiveness, and trustworthiness has become a critical research challenge.
The REACH-AI 2026 Workshop aims to bring together researchers, practitioners, policymakers, community stakeholders, and industry leaders to advance methodologies, datasets, benchmarks, and evaluation frameworks that enable culturally aware and socially responsible AI systems.
Topics of Interest
We invite original research, position papers, system demonstrations, datasets, benchmark papers, and case studies on topics including, but not limited to:
Cultural Competence in Generative AI
- Definitions, theories, and measurement of cultural competence in AI systems
- Cross-cultural and multilingual evaluation of LLMs
- Representation and inclusion of underrepresented communities
- Cultural adaptation and localization of foundation models
- Community-centered AI design methodologies
Evaluation and Benchmarking
- Novel benchmarks for culturally competent AI
- Human-centered evaluation methodologies
- Participatory and community-informed assessment frameworks
- Metrics for fairness, equity, inclusiveness, and social impact
- Benchmarking across demographic, linguistic, and geographic populations
- Evaluation of culturally grounded reasoning and contextual understanding
Knowledge Management and Information Access
- Culturally aware information retrieval and recommendation systems
- Knowledge representation of cultural and indigenous knowledge
- Retrieval-augmented generation for diverse communities
- Community-specific knowledge graphs and knowledge bases
- Cross-cultural information seeking and knowledge discovery
- Information retrieval for culturally relevant content via LLMs
- Knowledge construction for underrepresented communities
- Knowledge representation for culturally sensitive contexts
- Human-AI collaboration in knowledge acquisition
Trustworthy and Responsible AI
- Bias detection, mitigation, and auditing
- Transparency, explainability, and accountability
- Safety and alignment across cultural contexts
- Fairness in multilingual and multimodal foundation models
- Privacy-preserving and ethical evaluation practices
- Cultural/social bias mitigation for LLMs and multimodal models
- Human-centered AI evaluation and annotation frameworks
- Explainability and transparency in AI evaluation
Community Engagement and Applications
- Participatory AI development and co-design
- AI for education, healthcare, social services, and public policy
- Community-driven datasets and benchmark creation
- Case studies involving historically underserved populations
- Real-world deployment experiences and lessons learned
Submission Categories
We welcome the following types of submissions:
- Full Research Papers (up to 8 pages + references)
- Short Papers / Position Papers (up to 4 pages + references)
- Demo and System Papers (up to 4 pages + references)
- Dataset and Benchmark Papers (up to 6 pages + references)
Submissions should follow the ACM conference format and be submitted in PDF format through the workshop submission site.
Where to Submit: Submissions will be managed through the EasyChair system. Submission portal: HERE
Submission Guidelines
Authors are invited to submit original short papers that have not been previously published, and are not being considered for publication in any other forum. Authors of each submission to REACH-AI 2026 are required to nominate at least one of the authors of each submission, for reviewing some papers. This can be done in EasyChair. Note that failure to nominate a reviewer among the authors will lead to desk rejection of the paper.
- Full Research Papers: Manuscripts should be submitted using the ACM sigconf template, see https://www.acm.org/publications/proceedings-template. Submissions should be in 2-column sigconf format. Full papers cannot exceed 8 pages, including an appendix, plus unlimited pages for the GenAI Usage Disclosure section and references (paper content is limited to 8 pages, which means that if you have an appendix, then it should be included within that page limit. It is also promising if you do not have an appendix and instead 8 pages of content). Suitable LaTeX, Word, and Overleaf templates are available on the ACM website. Please use the sigconf proceedings template for LaTeX and the Interim Template for Word. ACM CCS concepts and keywords are required for review. For LaTeX submissions, please use the following document class: \documentclass[sigconf,natbib=true,anonymous=true]{acmart}. The review of full papers will be double-blind, and those submissions that have not been properly anonymized will be desk-rejected without review. At least one author of each accepted paper must register to present the work on-site in Rome, Italy as scheduled in the conference program.
- Short Papers / Position Papers: Manuscripts should be submitted using the ACM sigconf template, see https://www.acm.org/publications/proceedings-template. Submissions should be in 2-column sigconf format. Short papers cannot exceed 4 pages, including an appendix, plus unlimited pages for the GenAI Usage Disclosure section and references (paper content is limited to 4 pages, which means that if you have an appendix, then it should be included within that page limit. It is also promising if you do not have an appendix and instead 4 pages of content). Suitable LaTeX, Word, and Overleaf templates are available on the ACM website. Please use the sigconf proceedings template for LaTeX and the Interim Template for Word. ACM CCS concepts and keywords are required for review. For LaTeX submissions, please use the following document class: \documentclass[sigconf,natbib=true,anonymous=true]{acmart}. The review of short papers will be double-blind, and those submissions that have not been properly anonymized will be desk-rejected without review. At least one author of each accepted paper must register to present the work on-site in Rome, Italy as scheduled in the conference program.
- Demo and System Papers: Manuscripts should be submitted using the ACM sigconf template, see https://www.acm.org/publications/proceedings-template. Submissions should be in 2-column sigconf format. Demo and System papers cannot exceed 4 pages, including an appendix, plus unlimited pages for the GenAI Usage Disclosure section and references (paper content is limited to 4 pages, which means that if you have an appendix, then it should be included within that page limit. It is also ok if you do not have an appendix and instead 4 pages of content). Suitable LaTeX, Word, and Overleaf templates are available on the ACM website. Please use the sigconf proceedings template for LaTeX and the Interim Template for Word. ACM CCS concepts and keywords are required for review. For LaTeX submissions, please use the following document class: \documentclass[sigconf,natbib=true,anonymous=true]{acmart}. The review of Demo and System papers will be double-blind, and those submissions that have not been properly anonymized will be desk-rejected without review. At least one author of each accepted paper must register to present the work on-site in Rome, Italy as scheduled in the conference program.
- Dataset and Benchmark Papers: Manuscripts should be submitted using the ACM sigconf template, see https://www.acm.org/publications/proceedings-template. Submissions should be in 2-column sigconf format. Dataset and Benchmark papers cannot exceed 6 pages, including an appendix, plus unlimited pages for the GenAI Usage Disclosure section and references (paper content is limited to 6 pages, which means that if you have an appendix, then it should be included within that page limit. It is also ok if you do not have an appendix and instead 6 pages of content). Suitable LaTeX, Word, and Overleaf templates are available on the ACM website. Please use the sigconf proceedings template for LaTeX and the Interim Template for Word. ACM CCS concepts and keywords are required for review. For LaTeX submissions, please use the following document class: \documentclass[sigconf,natbib=true,anonymous=true]{acmart}. The review of Dataset and Benchmark papers will be double-blind, and those submissions that have not been properly anonymized will be desk-rejected without review. At least one author of each accepted paper must register to present the work on-site in Rome, Italy as scheduled in the conference program.
Review Process
All submissions will undergo peer review. Accepted papers will be selected based on originality, technical quality, relevance, significance, and potential to stimulate discussion within the workshop community.
Evaluation criteria include:
- Novelty
- Technical quality
- Relevance to workshop themes
- Potential impact
- Reproducibility and clarity
Each submission will receive at least three reviews.
Important Dates
(Tentative — to be finalized upon workshop approval)
- Paper Submission Deadline August 24, 2026
- Notification of Acceptance September 23, 2026
- Camera-Ready Papers Due October 7, 2026
- Workshop Date November 8, 2026