Threat Modeling in the Age of AI - Global Meetup Recap

On 2024-11-14T05:00:00Z, we hosted the November Global Community Meetup with Susanna Cox, where she highlighted the critical role of data in AIML security and helped us apply that knowledge to understand AI-specific threats using three methodologies.

Replay

Key takeaways

Susanna introduced three approaches to understanding AI-specific threats:

She also suggested three steps to understanding the AIML system attack surfaces, along with three questions you can ask in the process. Here’s a glimpse of what she covered:






:point_right: View the full slides.

Resources

Here are a few resources Susanna referred to during her session:

Peer discussions

After Susanna’s presentation, we moved into breakout groups groups to discuss:

  • How does the role of data in AI systems change the attack surface? How can security professionals adjust their thought processes around this?
  • Does the CIA triad (confidentiality, integrity, and availability) still apply to AIML systems? Why or why not?

:point_right: Take a look at the insights here.

Group photo :heart_eyes:

It was a really insightful session, packed with helpful frameworks and resources for modeling threats in any AIML environment - led by a true AI security pioneer. Thanks to everyone who joined us, and a special shoutout to our peer group facilitators @aulong @ClaireAA @Robin @dimitri.redant @FraserScott for guiding the breakout sessions!

Next global meetup (Jan 13, 2025)

We’ll dive into “success metrics for threat modeling” at our next global meetup, featuring Caroline Wong as our monthly speaker. Caroline wrote an entire book on metrics for software security and will join us to discuss what success looks like for threat modeling and how to measure and communicate them. Join us: TMC January Global Meetup: Success and Metrics for Threat Modeling · Zoom · Luma

2 Likes