The potential for an AI system or associated process to cause harm, produce unintended outcomes, or fail to achieve intended goals. Risk in AI encompasses technical failure modes, bias, adversarial vulnerabilities, misuse, and broader societal harms; risk management frameworks characterise likelihood and severity to prioritise mitigations.
Semantic Classification
Content
Risk — content pending enrichment.