Lessons Learned from Developing Secure AI Workflows at Google - RSA 2024

[@theodoreniedzialkowski3221]

Great resource, thank you for sharing! Have you considered something like an OWASP or CSA "Threat Modeling Toolkit" to use when evaluating attacks, capabilities, and mitigations to AI implementations? I think Google's SAIF and the way you broke it down by Data, Infra, Model, and App is a great start.