Talk With an Expert

The Future of AI Security Needs Your Voice 

The Critical AI Security Guidelines draft is now available! This first-of-its-kind framework provides expert insights into securing AI deployments, balancing security and scalability, and aligning with evolving governance and compliance requirements.

Key Takeaways from the Guidelines:

  • Secure AI deployments with a multi-layered security approach
  • Protect against model poisoning, prompt injection, and adversarial attacks
  • Implement governance frameworks that adapt to AI advancements
  • Balance security and scalability in AI model hosting
  • The biggest risk in AI? Not using it at all due to security concerns

Public comments are now open. We welcome your thoughts to help shape AI security standards! Learn how to contribute.

Download Today

Why Your Contribution Matters

  • Stay current. AI evolves daily—your real-world fixes keep us fresh.
  • Boost accuracy. Spot a typo or missing control? A quick PR stops confusion.
  • Share expertise. Your diagrams and war stories turn theory into practice.
  • Align with standards. Link to industry frameworks like OWASP AI Security so we all stay on the same page.
  • Build community. Every voice strengthens our shared guidelines.
  • Be recognized for your contributions.

How to Contribute

We welcome your fixes, clarifications, and examples. Here’s the simplest way to pitch in:

  1. Sign up on GitHub if you don’t have an account yet.
  2. Clone https://github.com/sans-community/ai-guidelines to grab the latest Markdown.
  3. Branch off using your last name plus month/year (for example: bromiley-may2025).
  4. Edit: add or tweak text, images, diagrams—and please include a direct link to the OWASP AI Security Framework (https://owasp.org/www-project-ai-security-and-privacy-guide/) wherever you reference controls or best practices.
  5. Push your branch and open a Pull Request.
  6. Review: we’ll drop feedback, help refine, and merge your contribution.
  7. Ask questions in our Community Chat or raise an issue in the repo.