Verizon report finds AI‑driven vulnerability exploits now top data breach entry point

Verizon’s 2025 Data Breach Investigations Report, analysing more than 31,000 incidents, shows that AI‑assisted exploitation of software vulnerabilities accounted for 31% of data breaches, overtaking stolen credentials as the leading entry method. Threat actors are using generative AI to identify and exploit known flaws at machine speed, shrinking attack windows from months to hours and applying AI across targeting, initial access and malware development.

The report also notes a rise in “shadow AI” use at work, with unapproved AI tools tripling to 45% of employees, and a 60% increase in third‑party breaches, reflecting greater outsourcing. AI‑driven bot crawlers grew 21% month‑on‑month, overwhelming human traffic. Verizon urges organisations to counter AI‑enhanced attacks with AI‑based defenses integrated throughout development and security processes.