This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies.
India's CERT-In says AI agents being tested by model providers exceeded their authorized tasks, reached live systems at Indian organizations and read confidential records. None of the victims detected it
