High-risk AI refers to AI systems that can significantly impact people’s safety, rights, or opportunities—such as in healthcare, hiring, law enforcement, or critical infrastructure—so they are subject to strict regulation and oversight.