An AI browser is a web browser that uses a model to interpret, summarize, search, or even act on web content on behalf of the user. Instead of only rendering pages, it can read text, follow links, fill forms, query connected services, and make decisions based on what it sees. That makes it more powerful than a conventional browser, but it also expands the attack surface.
In cyber security, the main risk is prompt injection: attacker-written content can be embedded in a page, document, or hidden element and treated like instructions for the browser agent. If the system cannot cleanly separate trusted user intent from untrusted web data, it may take actions the user never meant to approve. Defenses include strict instruction boundaries, least privilege, sandboxing, and explicit confirmation before sensitive actions such as sending data or changing account settings.