FTC Probes OpenAI, Anthropic as AI Agent Safety Risks Draw Scrutiny

The feed covers active exploitation of multiple critical vulnerabilities, including Roundcube, WordPress, Google Pixel, and iPhone flaws, as well as widespread browser security issues. It also highlights phishing campaigns targeting Microsoft 365 credentials, exposed surveillance-camera data, autonomous AI agent safety and privacy risks, and emerging security governance concerns. Organizations should prioritize patching vulnerabilities confirmed as exploited, harden identity controls against phishing, and assess the privacy and authorization boundaries of AI-enabled systems.