IBM: AI-driven attacks increased 56% last year, and data breach costs are up 12%

AI-based attacks that rely on deepfake impersonation and AI-enabled malware are getting faster and cheaper to launch, which has driven the cost of finding and fixing enterprise data breaches to a record high.
AI-enabled breaches cost an average of $6 million, which is roughly $1 million more than the global breach average of $4.99 million, according to IBM’s 2026 Cost of a Data Breach Report. The 2026 report, conducted by Ponemon Institute and sponsored and analyzed by IBM, is based on breaches experienced by 602 organizations globally between March 2025 and February 2026.
At $4.99 million, the global average cost of a data breach represents a 12% increase over last year’s numbers and a new high. That increase was largely driven by detection, escalation, and lost business costs, according to IBM. One in four malicious breaches were AI-enabled, IBM found.
AI is accelerating the attack lifecycle and changing breach economics, notes Limor Kessem, global lead, X-Force cyber crisis management at IBM, in a blog post about the report.
“Looking at the changes from last year’s report, AI-driven attacks increased by 56%, adding an average of $1 million per breach, as attackers use AI tools to increase speed, scale, and precision. This is not simply an evolution in attacker tooling; it is a structural shift,” Kessem wrote. “When adversaries can automate reconnaissance, generate persuasive phishing content, adapt malware and test exploits at machine speed, the cost and complexity of launching sophisticated attacks drops materially. Breaches become faster, broader and more expensive.”
“When attack velocity increases, the enterprise has less time to detect, validate and contain an incident. That compressed response window directly drives higher losses, whether through operational disruption, data exposure, legal costs, customer remediation or reputational damage,” Kessem continued. “From the report’s findings, two cost categories, detection and escalation alongside lost…