Block all AI browsers for the foreseeable future: Gartner

Gartner cautions that AI-enabled browsers and sidebars can expose sensitive web content and histories to cloud back ends and are vulnerable to agentic failure modes (e.g., indirect prompt injection, erroneous autonomous actions, and phishing-induced credential abuse). The analysts recommend rigorous risk assessments of back-end AI services, centralized privacy/security hardening, user education, and strict capability limits (such as disabling email and data retention), advising organizations to block AI browsers if risks are unacceptable and to monitor for prohibited use cases.