In many jurisdictions, open government and access to information laws such as Freedom of Information (FOIA) and Access to Information (ATI) require large-scale public disclosure of government records, resulting in massive, multimodal data collections whose complexity increasingly challenges both legal compliance and technical processing. At the same time, governments face strict legal obligations to disclose information within statutory deadlines, while protecting sensitive and personal information.
The workshop addresses two key perspectives:
- AI for citizens: Tools and techniques for improving search, exploration, and understanding of public government information
- AI for governments: Assisting in accessibility, pre-processing, metadata enrichment, retrieval, filtering, and protecting sensitive information consistent with public disclosure laws
Topics of InterestTopics include, but are not limited to:
- AI-augmented search, retrieval, and summarization for public records
- Technology-assisted review for FOIA and records access requests
- Automated sensitivity review and redaction under FOIA/GDPR
- Metadata enrichment and entity extraction for government record discovery
- Multimodal processing of scans, PDFs, and legacy document formats in public archives
- Agentic AI for FOIA request triage and handling
- Public-facing tools for navigating heterogeneous government data repositories
- Formalising legal standards for disclosure, exemptions, and harm in AI-assisted access workflows
- Governance, auditability, and explainability of AI-assisted disclosure, including human-in-the-loop review
- Automated classification for government records retention
- Retrieval Augmented Generation (RAG) architectures and generative AI for public government records and cultural heritage archives
- AI-assisted declassification of government records for public releaseSubmission Guidelines
We invite submissions of:- Research papers (3-9 pages + references): original research contributions
- Position papers (up to 5 pages + references): insights from practice
Paper must be formatted using the ACM sigconf template (for LaTeX) or the interim template layout.docx (for Word), both at
http://www.acm.org/publications/proceedings-template. Papers should be submitted via the submission portal at
https://submit.aiog.net. Reviewing will be double-blind, i.e., papers submitted for review must not include names and affiliations of the authors. Accepted papers will be published in OpenReview proceedings.
Authors of accepted papers will be invited to present their work in person in Singapore.