Fresh Attack Reveals Yet Another Security Risk of AI Browsers

AI browsers offer convenience by executing complex tasks through simple prompts, but security researchers warn that current guardrails are reactive and insufficient. A new study demonstrates how malicious websites can trick AI browsers into a false reality where safety rules are ignored. This allows attackers to extract private code or steal credentials from password managers, highlighting a fundamental design flaw.