Insufficient standardization – AI laws differ throughout nations around the world and industries, making compliance complex. Robustness: AI-powered devices has to be actively defended from adversarial attacks, minimizing stability pitfalls and enabling self-assurance in system outcomes.