Agentic web browsers could be programmed to fall for phishing and scam traps by using artificial intelligence (AI) capabilities to carry out actions across several websites on behalf of a user.
According to a paper Guardio shared with The Hacker News prior to publication, the attack essentially exploits AI browsers’ propensity to rationalize their behavior and use it against the model itself to decrease its security barriers.
These days, the AI works in real time within chaotic and dynamic pages, constantly seeking information, making choices, and explaining its actions as it goes. “Narrating” is an understatement, according to security researcher Shaked Chen. “It blabbers, and way too much.”
This is what we refer to as “agentic blabbering,” where the AI browser reveals what it sees, what it thinks is going on, what it intends to do next, and what signals it deems safe or suspicious read more about Researchers Trick Perplexity’s Comet AI Browser Into Phishing Scam in Under Four Minutes.
Get up to date on the latest cybersecurity news and enhance your knowledge of cybersecurity with our thorough coverage of the dangers, breaches, and solutions.