One In Four Breaches Are AI-Enabled, And That’s Before Hugging Face

IBM says 25% of malicious breaches involved AI, up 56% year over year, with AI-enabled incidents averaging about $6 million in cost.

The report suggests frontier AI is making phishing, malware, and deepfake impersonation cheaper and more effective for attackers.

At the same time, enterprises are increasing security spending, even as AI tools can help defenders reduce breach costs.

Separately, Veracode found that more than 40% of code generated by tested LLMs contained vulnerabilities, highlighting risks in AI-assisted development.