Scientists invented a fake disease. AI told people it was real

Develop an adversarial framework to test LLM susceptibility to fabricated disinformation. This tool would help security teams audit model grounding before deployment.

Suggested repo: hallucination-stress

"Test if your AI believes in fake realities."

Estimated effort: 60h