The ETSI Securing Artificial Intelligence (SAI) group was launched to address three critical areas: defending AI systems from attacks, mitigating AI used as an attack vector, and leveraging AI to enhance security measures. The group will initiate three key activities, starting with an AI Threat Ontology to standardize terminology for threats against AI systems, attackers, and defenses. A second activity will produce a Securing AI Problem Statement to prioritize potential threats and guide security-by-design principles, while a third will deliver a Data Supply Chain Report analyzing methods to ensure data integrity and confidentiality for AI training. Founding members include BT, Huawei Technologies, NCSC, and Telefónica, with the inaugural meeting scheduled for October 23 in Sophia Antipolis.
Keywords: AI threat ontology, data supply chain integrity, autonomous system security, AI attack vectors, secure AI development