Overview
- The IBM Cost of a Data Breach Report, released July 29, 2026, found 25% of malicious breaches were AI-enabled, a 56% rise from the prior year and about $1 million more expensive than the overall average.
- AI-driven incidents averaged roughly $6.0 million per breach while the global mean breach cost rose to $4.99 million, driven by longer detection and containment times that now average 247 days.
- Attackers are using deepfake impersonation and AI-enabled malware and often exploit weak surrounding systems such as compromised APIs and cloud misconfigurations that let them reach models and apps.
- Organizations that deploy AI and automation across prevention, detection and response close breaches about two months faster and save close to $2 million per incident, yet 92% of firms hit through AI systems lacked basic access controls.
- The report warns of concentrated systemic risk with 62% of AI-driven attacks hitting critical infrastructure sectors and many firms showing gaps in encryption, cryptographic visibility, vulnerability management and shadow AI use, prompting 85% of surveyed organizations to plan higher security budgets.