When AI Fails, What Works? A Data-Driven Taxonomy of Real-World AI Risk Mitigation Strategies

Evgenija Popchanovska, Ana Gjorgjevikj, Maryan Rizinski +3

Large language models (LLMs) are increasingly embedded in high-stakes workflows, where failures propagate beyond isolated model errors into systemic breakdowns that can lead to leg…