Now you can “wee what matters most,” Google told Gmail users last week. With the new AI inbox first announced in January, users can essentially put Gemini in charge to “to help you better manage and stay on top of your inbox.”

“Today, 3 billion users rely on Gmail to connect and get things done,” Google says. But email is a comms platform. Getting things done may involve sending emails, but until now the inbox hasn’t actually been “getting things done” itself.

That all now changes. “Today, we’re rolling out new capabilities that let you take action and get things done seamlessly,” Google explains.

An AI Inbox doesn’t just receive emails, it “personalizes draft replies.” That means by the time you read an email, Gmail has already generated “a contextual draft so you can review and respond in seconds." Maybe even without reading it.

Similarly, why bother searching your own files when Gmail can do it for you. “No more digging through threads. When a task requires reviewing a Google Doc, Sheet or Slide, the relevant link is surfaced right next to your to-do.”

Google assures that “this analysis happens securely with the privacy protections you expect from Google, keeping your data under your control,” but I’m sure you can see the problem . Beyond the indirect prompt injection risks when AI reads emails and writes replies before you can, there’s a privacy shaped hole .

I’m a security and privacy writer. That’s where I go first with all these changes. I don’t doubt the convenience and time savings. An inbox that manages itself sounds great. But these changes are being rolled out with little debate as to the pros and cons. Most are enabled by default. Users can opt out, but most won’t .

So, give it some thought. Use the new features by all means. But Google’s Gemini is now encroaching into emails and calendars, photos and file storage. At some point, it might make sense to ask how much of your data you want to open up.