Full papers
Submissions limited to 4 pages for main contents with unlimited references and appendix pages; accepted papers may include 1 additional page (5 pages total) for camera-ready version.
NLP/CORE Unranked
SeT-LLM @ KDD 2026
KDD Workshop on Secure and Trustworthy Large Language Models
SeT-LLM @ KDD 2026 is a workshop focused on addressing security and trustworthiness challenges in Large Language Models (LLMs) as they are deployed in real-world, data-centric systems. It brings together researchers and practitioners to explore solutions across theory, algorithms, and applications, including reliability, privacy, safety, and robustness in LLM-powered systems.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Research areas in scope
01
Security
prompt injectionjailbreaksadversarial promptingmodel stealingunsafe tool userobust defense mechanisms in LLM development and deployment
02
Trustworthy Assurance and Evaluation
benchmarksauditinguncertainty estimationcalibrationrisk-sensitive evaluation methodologies
03
Privacy and Data Protection
memorizationprivacy leakagedata governancesecure fine-tuningprivacy-preserving inference
04
Robustness and Reliability
Distribution shiftlong-context failureretrieval robustnessdependable reasoningresilient multi-agent behavior
05
Alignment and Safety
harm mitigationpolicy compliancevalue alignmentcontrollabilitysafe deployment strategies
06
Applications and Case Studies
Healthcarescienceeducationlegal reasoningpublic-sector useenterprise systemsother high-stakes settings
07
LLM-based Agents Security & Safety
security vulnerabilities in agent tool-usemulti-agent cooperation risksautonomous decision-making alignmentsafety guardrails for LLM agents in open-world environments
Policies worth checking twice
- SeT-LLM '26 follows a double-blind review process.
- Submitted manuscripts must be anonymized with identifying information removed.
- Submissions concurrently under review at other venues are acceptable.
- All accepted papers are non-archival and will be made publicly available on the workshop website without an official proceeding.
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026