Claude can send emails without asking; Google still won't let Gemini press send on its own

Anthropic’s Claude can now autonomously send Gmail messages, while Google’s Gemini still requires a human click, highlighting a strategic split between aggressive agentic automation and platform-level caution. For CIOs and technology leaders, the business upside is clear: AI can materially improve productivity in high-volume communications, but autonomous sending also raises governance, compliance, and security risks, especially around prompt injection, brand exposure, and irreversible errors. IT organizations will need stronger policy controls, approval workflows, auditability, and use-case segmentation so they can capture efficiency gains without expanding operational or reputational risk.