In recent research, AI browsers found vulnerable to hidden attacks.
AI browsers can be manipulated through hidden commands.
Hackers hide secret instructions in invisible website text.
In Brave’s test, clicking “Summarize this page” triggered a full data leak.
The AI fetched emails, passwords, and even posted them online.
Those AI browsers run with your logged-in privileges.
The AI executes hidden commands — without you knowing.
AI in your browser is powerful — and dangerously obedient.