The road to trustworthy AI: Strategies for robustness