2025-11-07 –, Room 334
Prepare for incidents with the AI Incident Response & Control (AIRCTL) Project maintainers. You will learn about AI risks, why preparedness matters, and incident response (IR) scenarios using FOSS resources.
We will guide you through how to run an in-person or virtual tabletop role-playing session using an AI-themed expansion to the game Backdoors and Breaches from Black Hills Information Security. All tools for gameplay and resources are included and FOSS.
Attendees will leave with all the practical skills needed to run a realistic tabletop IR exercise for safeguarding an organization using AI technology and for protecting research & development (R&D) assets. NO artificial intelligence or incident response experience needed!
After learning how to be incident commanders, you will immediately apply the skills learned in-session to run tabletops at SeaGL with instructor-led facilitation.
Participants of will attend an introductory briefing by the AIRCTL team on AI for security professionals, four common types of AI incidents, and the mechanics of the Backdoors & Breaches tabletop IR game.
Participants will then form teams, get guidance for your scenario, and breakout to roleplay as an Incident Responder or Incident Commander for a facilitated playtest of one of the common AI incident scenarios: R&D compromise, material breach via AI path, AI defacement & manipulation, AI exploitation.
Be prepared to commit at least one hour to the workshop session to experience all segments. The intro briefing is 10 minutes with short Q&A. Each game scenario takes 30-60 minutes for play, depending on the IR team. Debrief discussion and exit survey will take roughly 20 minutes. An optional “all hands” debrief will follow.
Participants who playtest any of the AIRCTL Project IR scenarios and share their experiences and ideas as feedback will be eligible for playtester credit on the AIRCTL project pages — we want to see the community win and appreciate your time with us has value. The session will not be recorded but notes and photos with consent are welcome.
None
Emily Soward has over 15 years of experience in AI R&D with specializations in governance, operations, and security. She is a serial innovator, founder, and leader in AI incident response, ecological and edge AI research, and AI security research. She is notable for courses and teaching in AI governance, risk management, operations, and security, as well as her work on AI and ML frameworks for Amazon Web Services. Her contributions to the HITRUST Alliance AI Working Group supported the launch of the first cybersecurity certification for deployed AI systems. In 2024, she cofounded the AI Incident Response & Control (AIRCTL) Project, an open-source project making top AI security skills accessible for small and under-resourced organizations.