The technical robustness of high-risk AI systems should include robustness in terms of accuracy and resilience to attempts by unauthorised third parties to alter their use, outputs or performance by exploiting the system's vulnerabilities. Technical solutions to address AI-specific vulnerabilities could include technical measures aimed at preventing, detecting, responding, resolving and controlling attacks trying to manipulate the training data set (data poisoning), or pre-trained components used in training (model poisoning), inputs designed to cause the AI model to make a mistake (adversarial examples or model evasion), confidentiality attacks or model flaws. Reference should be made to Regulation (EU) 2019/881 as regards certain key aspects of cybersecurity, including cyber resilience and the prevention and addressing of intentional cyber incidents such as cyberattacks.