Managing the Trust-Risk Equation in AI: Predicting Hallucinations Before They Strike Security Week News