Bedrohungs-Briefing: Einblick in den GenAI-Angriffslebenszyklus

In this episode of Threat Briefing, we dive into the real-world tactics adversaries use to exploit generative AI systems. Security researcher Alex Groyz joins us to break down how attackers are actively abusing cloud-based LLM infrastructure, from exposed inference endpoints and Shadow AI to prompt injection, jailbreaks, and inference-driven data exfiltration. We explore how the MITRE ATLAS framework maps to these attacks and discuss how defenders can use it to simulate adversarial behavior, validate detections, and secure GenAI across the full lifecycle.