Lately I’ve seen this phrase Good AI vs Bad AI, a lot in cybersecurity reporting. Defensive AI (think anomaly detection, predictive threat modeling, self-healing networks) is stacking up against offensive AI (malware that evolves, AI-powered phishing, deepfakes, etc.).  

At the same time, debates from Black Hat and DEF CON are spotlighting how AI tools for defenders are gaining traction, but so are AI tools for attackers leveraging open-source LLMs. 

From a learning perspective, I’m trying to wrap my head around how to train defensive models effectively when the threat models themselves are AI-driven. I’ve been exploring Haxorplus for guided content on designing secure AI and understanding adversarial scenarios alongside general ML platforms like Kaggle or academic labs.

Would love to crowdsource ideas: how are you guys bridging that gap?