AI-infused web browsers are here and they’re one of the hottest products in Silicon Valley. But there’s a catch: Experts and the developers of the products warn that the browsers are vulnerable to a type of simple hack.

The browsers formally arrived this month, with both Perplexity AI and ChatGPT developer OpenAI releasing their versions and pitching them as the new frontier of consumer artificial intelligence. They allow users to surf the web with a built-in bot companion, called an agent, that can do a range of time-saving tasks: summarizing a webpage, making a shopping list, drafting a social media post or sending out emails.

But fully embracing it means giving AI agents access to sensitive accounts that most people would not give to another human being, like their email or bank a

See Full Page