Lessons From AI Hacking: Every Model, Every Layer Is Risky

https://www.darkreading.com/application-security/lessons-ai-hacking-model-every-layer-risky