Google wants you to worry about hackers reading your email while it trains artificial intelligence on your private conversations—and the tech press is playing along.
Engadget published a guide this week to "securing" your Gmail account, walking readers through two-step verification, passkeys, and checking for suspicious forwarding rules. All useful against intruders. What Engadget never breathes a word about: the biggest entity already reading every word in your inbox isn't some phantom attacker. It's Google.
Meanwhile, 9to5Google reports that Google is rolling out new Gemini app settings that finally let users keep their chat history without letting the company train AI models on their content. Until now, Google's "Keep Activity" setting bundled those choices together—save your history, or let us use it to build our AI. Pick one. The new update splits them: one setting controls how long chats are saved, another controls whether Google feeds your words into its generative AI training pipeline. If you currently have Keep Activity turned off, both new settings will default to off. If you haven't touched it, Google has been helping itself.
This is the pattern. Google builds surveillance into the default. Then it offers buried settings menus to opt out. Then the tech press writes "helpful" guides that treat the surveillance itself as a non-issue.
Here is what you actually do—on Google's terms, since it won't give you better ones.
Lock the front door. Turn on two-step verification. Go to Security & Sign-in in your Google Account, find "How you sign in to Google," and enable it. A stolen password alone will not grant access; sign-in also requires something physical, like your phone.
Set up passkeys. Passkeys use your fingerprint, face scan, or screen lock instead of a password and resist phishing. Go to your account's passkey page and select "Create a passkey." Only do this on devices you control—anyone who can unlock that device gets into your account.
Audit who has access. Go to myaccount.google.com, open Security & Sign-in, scroll to linked apps, and select "See all linked apps." Remove anything you don't recognize or no longer use.
Check Gmail for quiet backdoors. Open Gmail on desktop, click the gear icon, then "See all settings." Check automatic forwarding, filters, labels, scheduled emails, vacation responder, and IMAP/POP access. Under Accounts and Import, look for delegated access you didn't authorize. Automatic forwarding is the nastiest—it silently copies every new message to another address. Engadget flags this: if you find a forwarding rule you didn't set up, change your password immediately and kill the rule.
Kill AI training on your Gemini chats. When the new settings roll out over the coming weeks, turn off the option that lets Google use your chat content—including text and attachments—to train its generative AI models. You can keep your history; you don't have to feed the machine.
Google offers these controls because it knows most people will never find them. Every step is an opt-out, buried in menus, framed as a "recommendation" instead of a right. The company that reads your mail, scans your attachments, and trains its AI on your conversations tells you to check for suspicious forwarding rules. The irony writes itself.
The open question isn't whether these settings work. It's why the default is always surveillance—and why the press that claims to inform you won't name the watcher.







