Technical
Model robustness
Resistance of AI models to errors, attacks and unexpected inputs.
Definition
A technical characteristic of AI systems referring to resilience against adversarial manipulation, distributional shifts and operational failures; highlighted as a cross‑cutting requirement alongside cybersecurity, testing and auditability.
Related Terms
safety and controllability
Ensuring AI is transparent, robust, and traceable....
Reliable AI (Trustworthy AI)
AI systems meeting ethical, legal, and technical requirements....
Model security
Measures to protect models from theft, tampering, or misuse....
Security (AI)
Protection of AI systems from malicious attacks....
Auditability of the AI system
Capacity to evaluate algorithms, data, and design or undergo audits....