Malwarebytes reports rise in AI-assisted cyberattacks in 2025

AI use in cybercrime has escalated last year, with the technology prompting accelerated and more effective cyber intrusions, reports Cybersecurity Dive.

Despite the emergence of the first AI-driven intrusions, including deepfake-based social engineering, hands-on-keyboard attacks continued to be the leading incidents in 2025, according to a Malwarebytes report. Increased focus on Model Context Protocol exploitation in AI attacks has also been urged. Additional findings also revealed intensified ransomware activity, with overall attacks increasing 8% year-over-year to record highs and 86% of intrusions involving remote encryption, often launched from a single unprotected system. The report added that the United States accounted for 48% of detected ransomware attacks, with incidents spanning 135 countries, highlighting how AI-enabled techniques are spreading across the global cybercrime landscape.

"This pattern reflects long-standing geopolitical and economic dynamics in the ransomware ecosystem: Cybercriminals focus on wealthier economies with familiar technology stacks and languages, and where political or law-enforcement blowback is minimal," the company emphasized.