Browser-Based Attacks

Browser-based attacks are threats that target users or AI tools through the web browser, such as malicious web pages, hidden instructions planted in sites that an AI browsing agent reads, or scripts that steal logins and session data. For enterprise AI governance, they matter because AI assistants and agentic browsers can be tricked into taking actions on a user's behalf, like revealing confidential information or submitting forms, without the user noticing. Compliance and risk teams need to know how these attacks reach employees and what controls, such as restricting which sites AI tools can access, limit exposure.

1 item

ResearchGlobal2026-10-22

AI Tool Lures Drive 67% of Browser-Based Attacks, Exposing Shadow AI Controls

Research from Push Security shows ClickFix attacks reached 67% of browser-based attack detections in August 2026, up from 52% in the first half of the year. Attackers now use fake install pages for AI tools including Claude, ChatGPT, and Codex as their primary lure. Employees searching for unapproved AI tools through search engines are the primary exposure path.

Claude ChatGPT shadow IT AI tool impersonation browser-based attacks supply chain risk credential theft acceptable use