I Forced an AI to Give Me Its Password | Prompt Injection 101

A hands-on walkthrough of prompt injection attacks against AI systems, using the Gandalf AI lab platform across multiple difficulty levels. Covers techniques like indirect questioning, encoding tricks (binary, Base64, dashes), and multi-stage prompt building to bypass AI security filters including reflection models and GPT-based output filters. Also explains why AI integration into company infrastructure creates new attack surfaces and how these vulnerabilities relate to OWASP's top 10 LLM risks.