AI-Enabled Cyberattacks Rose 89% in 2025

According to CrowdStrike’s 2026 Global Threat Report, AI-enabled adversaries increased their operations by 89% in 2025. In its 2026 Cost of a Data Breach study, IBM reported that one in four security breaches is now AI-enabled. These AI-enabled breaches not only grew in number, but also caused higher damages, costing around $6 million on average against a $4.99 million global average.

Frontier AI labs are unable to manage risks from AI-enabled attacks by themselves. In July, when an OpenAI model broke out of a test environment and accessed Hugging Face servers, OpenAI relied on CrowdStrike to verify the incident, as I/O Fund reported in September.

Companies are also responding in a similar manner, as awareness about the potential of frontier models in cyberattacks grows. According to IBM, 85% of organizations plan to raise security spending after learning about frontier models’ cyber capabilities, up from 64% in 2025.