High-Risk AI System
An AI system whose failure or malfunction could reasonably cause significant harm to individuals' health, safety, fundamental rights or important societal interests, particularly in critical infrastructure, employment, education, law enforcement or benefit allocation.
In plain language
AI that could seriously hurt people if it makes mistakes. Examples include AI used in hiring decisions, medical diagnoses, bank loan decisions or police investigations. These systems face stricter rules because the stakes are high.
Why this matters
Classifying your AI systems by risk level is essential governance work. Understanding which applications qualify as high-risk determines the level of transparency, testing, documentation and human oversight required under emerging Australian and international regulations.
Relevance
GovernanceRisk classification determines regulatory obligations and the governance and oversight mechanisms your organisation must implement.
Related terms
Putting high-risk ai system into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.