Live Cybersecurity Training Session Highlights AI Security and Prompt Injection Challenges

The video features a live cybersecurity training session hosted by John Hammond of Just Hacking Training, focusing on AI security and prompt injection challenges. Andrew Bellini, creator of Only Lands, demonstrates a free interactive lab at onlylands.ai that teaches prompt injection techniques through gamified levels, including a new router lab simulating command execution on a Linux-based network device. The challenges use AWS Bedrock for AI models (Llama for Level 1, Amazon Nova for Level 2) and Docker containers with BusyBox for safe command execution, with Level 3 incorporating AWS classifiers to detect malicious inputs. A July 4th discount code (24250) for 20% off courses was active until July 31, while the team announced plans for a fall AI security course and additional challenges, such as agentic workflows and IoT hacking. Andrew also discussed local AI model experimentation, favoring AMD hardware (GMK Tech Evo X2) and the Quen 3.6 Mixture of Experts model with obliterated tuning for cybersecurity applications. The session highlighted upcoming appearances at DEF CON, including a free IoT hacking workshop in the IoT Village.