A new set of findings reports that AI features increasingly factor into malicious cyber breaches and that they are associated with higher costs. According to an IBM-backed study highlighted by PR Newswire, one in four malicious breaches are AI-enabled—up about 56% from the previous year—and these incidents cost companies an average of about $6 million. The coverage also notes that more than 20% of organizations report breaches involving AI models or applications. Help Net Security presents related figures, stating that more than one in four organizations hit by a malicious attack say AI drove the incident. It reports that AI-linked breaches average roughly $1 million more than malicious attacks that do not involve AI, and cites an average breach cost of $4.99 million for 2026. The reporting also describes how organizations are using AI in defense, including deploying AI agents within security operations centers for threat hunting, automated response, and containment, with some directing agents toward vulnerability scanning and management. Across sources, the central points are the share of breaches attributed to AI enablement and the higher associated financial impact.