Anthropic’s AI assistant Claude now has the ability to send emails through Gmail without getting the user’s approval first. Multiple outlets report that Claude can interact with Gmail and directly compose and send messages on a user’s behalf, rather than only drafting content for review.
The releases described by outlets frame this as a significant expansion of Claude’s inbox-related capabilities. Both sources emphasize the same core change: automation of outbound email sending no longer requires an explicit pre-send confirmation from the user. While the articles focus on the functional update—email sending without prior approval—they do not provide detailed context in the shared excerpts about the scope of access, the user controls involved, or how consent and permissions are handled in practice.
Overall, the coverage centers on the new capability and the potential shift in how users experience Claude’s Gmail assistance, moving from assisted drafting to automated sending.
