Google says Gemini breached three outside companies in a security test, guessing one password and using credentials found in a public repository for two others. Google said the model stopped after recognizing the targets were real; the guidance recommends long, unique passwords.
The article says AI models can guess more password variations in less time than human hackers. It advises avoiding common choices and information available publicly, using a different password for every service, mixing character types, and using a password manager to store and autofill credentials.
