High-risk AI systems

A category of AI systems to be classified in the law's executive by-laws that captures systems presenting significant risks (e.g., to health, safety, fundamental rights or bodily integrity); such systems will be subject to stricter governance, classification, and compliance requirements as detailed in the forthcoming by-laws.

AI systems identified as posing significant potential impacts on fundamental rights, safety, or other high-stakes outcomes; such systems are subject to enhanced pre-deployment risk assessment, registration, oversight, and compliance requirements under the bill.

AI systems whose deployment in sectors such as healthcare, education, justice, and employment could significantly affect individuals' rights, safety, or access to essential services; these systems are subject to rigorous pre-deployment evaluations and stricter governance measures under the strategy.

AI systems whose malfunction, bias, or erroneous outputs could create significant harm to individuals or public interests; explicit examples include medical diagnostic tools, autonomous vehicles, and systems used in the justice system. Such systems are subject to enhanced obligations including conformity assessment, registration, stricter documentation, and market surveillance.

AI systems that are specifically intended to autonomously make, or be a substantial factor in making, consequential decisions that have a material legal, or similarly significant, effect on the provision or denial to any consumer of critical services or opportunities. (While explicitly defined in other Virginia AI legislation like HB 2094, the Commission's mandate to prevent discrimination implicitly covers such systems).

High-risk AI systems are those identified by regulatory frameworks, such as the EU AI Act, as having the potential to cause significant harm to health, safety, or fundamental rights. These systems are subject to stringent requirements regarding conformity assessment, risk management, and market surveillance.

AI systems that pose a significant risk of harm to the fundamental rights, safety, or well-being of individuals, or to critical infrastructure, national security, or democratic institutions, triggering enhanced requirements for risk management, transparency, testing, and oversight. The determination is based on severity and probability of harm, context of use, and mitigation measures.

AI systems identified based on their potential to cause significant harm to health, safety, fundamental rights, or critical infrastructure, necessitating stricter regulatory oversight and compliance requirements.

High-risk AI systems are those identified under the AI Act as posing significant risks to fundamental rights and safety, necessitating stricter regulatory requirements. These systems are subject to rigorous conformity assessments, risk management systems, human oversight, and robust data governance to mitigate potential harms.

AI systems that could directly affect individuals' lives or rights in critical areas such as healthcare, justice, credit rating, public safety, or employment decisions, and are subject to stricter regulatory requirements.

AI systems identified in Annexes I and III of the AI Act that pose significant risks to health, safety, or fundamental rights, and are therefore subject to stringent requirements and obligations.

AI systems categorized by the EU AI Act as posing significant risks to health, safety, or fundamental rights. These include systems used in critical infrastructures, education, employment, law enforcement, and medical devices, and are subject to stringent requirements and conformity assessments.

High-risk AI systems are those classified as posing significant risks to the health, safety, or fundamental rights of individuals, as detailed in Annex III of the EU AI Act. These systems are subject to stringent requirements and conformity assessments due to their potential impact in critical sectors.

AI applications that pose significant risks to people's health, safety, or fundamental rights, and which are subject to stringent requirements and conformity assessments under the EU AI Act.

While not explicitly defined, the strategy acknowledges the need for future legislation to differentiate between various types of AI systems, their risk levels, and specific applications, mirroring international trends in AI governance. This implies a future classification of 'high-risk AI systems' that will be subject to more rigorous controls.

These systems are subject to stringent requirements under the EU AI Act, including conformity assessments, risk management systems, human oversight, and robust data governance. Nkom will primarily focus on market surveillance for these systems.

These systems are categorized into two types: AI used as safety components in critical infrastructures or products covered by EU product safety legislation, and AI used in sensitive areas like employment, education, law enforcement, migration, and justice. They require robust risk assessment, high-quality datasets, logging, documentation, human oversight, and cybersecurity.