High-risk system

An AI system or use-case identified as posing significant potential impacts on individuals' rights, safety, health or access to essential public services; such systems require heightened assessments, DPIAs, pre-deployment testing, human oversight, enhanced contractual protections and ongoing monitoring in line with EU risk-based rules.

A risk-category concept the decree directs the forthcoming General Law to define and regulate; it denotes systems whose deployment can materially affect fundamental rights, safety, or critical-sector operations and that therefore will be subject to enhanced obligations such as transparency, conformity assessment, human oversight and incident reporting.

AI systems classified as having material impact on individuals’ rights, safety, health, or fundamental interests (for example, healthcare diagnostics, credit scoring, or criminal justice applications), which are subject to scaled obligations such as AI impact assessments, enhanced testing, and registration.

An AI system or specific use-case identified under the bill's risk-classification framework as presenting substantial risks to fundamental rights, safety, health, or economic interests; high-risk systems are subject to pre-market conformity assessments or mandatory registration and enhanced obligations for providers and operators.

An AI system classified by the law as posing a substantial risk to fundamental rights or critical legal/administrative interests; such systems are subject to stricter obligations including impact assessments, registration on a public registry, periodic audits, and conformity assessment procedures.

An AI system classified by the bill as posing significant risk to safety, rights, or public interest; such systems trigger enhanced obligations including algorithmic impact assessments, conformity assessment, certification, and human-oversight requirements.

AI systems identified by the law as having the potential to cause significant harm to individuals, public interests, or sectoral functions; such systems are subject to mandatory risk classification, ethical/social/economic impact assessments, enhanced documentation, monitoring, and proportionate mitigation measures enforced by SGTD and sectoral supervisors.

Systems whose malfunction, biased outputs, or erroneous decisions have the potential to cause significant harm to individuals' rights, safety, or economic standing; such systems trigger mandatory risk assessments, DPIAs, independent testing and stronger oversight measures under the guidance.