Security & Privacy Foundations
for Large Language Models

About the Workshop

Why SPF-LLM?

Language models are rapidly becoming autonomous agents — orchestrating APIs, calling tools, executing multi-step plans, and collaborating with other models. This shift fundamentally changes their risk profile. Threats now span reasoning chains, tool interfaces, retrieval-augmented pipelines, and inter-agent communication, creating layered attack and privacy surfaces that no existing defense addresses comprehensively.

SPF-LLM creates a dedicated venue connecting three under-integrated communities: NLP/LLM researchers, security and privacy researchers, and production practitioners. We examine how model architecture, training, context handling, tool use, and multi-agent coordination create novel vulnerabilities — and how rigorous empirical and theoretical methods from NLP and security can close these gaps.

Key Deadlines

Important Dates

All deadlines are per the official AAAI-27 workshop calendar and are Anywhere on Earth (UTC-12). Dates subject to change — check back for updates.

Submissions

Call for Papers

We invite submissions on all aspects of security and privacy for large language models, with particular emphasis on agentic and multi-agent settings. We especially encourage works in progress, negative results, and cross-disciplinary perspectives.

Program

Workshop Schedule

The full schedule will be announced after notifications. The following is a tentative structure for the full one-day workshop.

Invited Speakers

Keynote & Panel Speakers

Catherine Régis

Full Professor, Faculty of Law, Université de Montréal (UdeM), Co-director of the Canadian AI Safety Institute research program and Director of Social Innovation and International Policy at IVADO

Panelist

Deep Learning, Online Learning

Ram Shankar Siva Kumar

Founder, Microsoft AI Red Team

Panelist

AI security, adversarial ML, Generative AI systems

Golnoosh Farnadi

Assistant Professor, HEC Montréal; Canada CIFAR AI Chair, Mila

Panelist

Deep Learning, Generative models

Konstantin Berlin

Cisco AI Defense - AI Detection Architect

Panelist

Organizing Committee

Workshop Organizers

Dr. Shanu Sushmita

Associate Professor, Northeastern University (Seattle)
Founder/Director, Generative AI Research Lab (main contact)

LLM safety, jailbreak detection, adversarial robustness, NLP

Dr. Mohammad Zandsalimy

Senior Research Engineer @Synopsys Inc.

Formal verification, AI security, mathematical encoding jailbreak attacks

Dr. Ehsan Aghaei

AI & Security Lead @Cisco

LLM safety evaluation, AI supply chain integrity, agentic reasoning pipelines

Dr. Giorgio Severi

Senior AI Safety Researcher, Microsoft AI Red Team

Adversarial ML, training-time adversarial interference, data poisoning

Dr. Mayana Pereira

Privacy Enhancing Technologies | Director at Capital One

Differential privacy, federated learning, responsible ML deployment

Dr. Martine De Cock

Professor, University of Washington Tacoma (directs the Privacy-Preserving ML Group)

Privacy-preserving ML, federated learning, trustworthy AI

Web Masters

Workshop Web Masters

Vivida Siddegowda Mallesh

Northeastern University (Seattle)

Kanchana Mala Belgam

Northeastern University (Seattle)

Program Committee

Interested in Reviewing?

We are looking for reviewers with expertise in LLM evaluation, NLP, ML systems, alignment, or production AI. If you'd like to contribute, fill out the form below — we'll be in touch.

FAQ

Frequently Asked Questions

Get in Touch

Contact & Information

Have questions about the workshop, submissions, or sponsorship inquiries? Reach out to the organizing committee.