Google doesn't ask. It assumes.
The company slipped a fundamental change into its privacy settings this June — no blog post, no banner, no press release. A customer email. That was the notice. Buried in that email: Google now claims the right to store and train on the photos you snap with Lens, the voice queries you speak to Search Live, the translations you practice aloud. Your media. Your voice. Your images. All of it, vacuumed into the training pipeline unless you hunt down the off switch.
This is not how a company that respects its users behaves. This is how a company that treats its users as training data behaves.
Two new settings appeared: Search Services History and Personalized Recommendations. They sound benign. They control what Google saves and how long it keeps your activity. But toggle "Save Media" on — the default — and suddenly every image uploaded to Lens, every audio clip sent to voice search, every spoken phrase in Translate becomes grist for Google's AI mills.
The scope is wider than Search. Maps. Shopping. Flights. Hotels. News. Translate. Any Google search surface that accepts media input now feeds the model. Snap a photo of a rash for Lens? Saved. Practice French pronunciation in Translate? Saved. Ask Maps for directions by voice? Saved.
Google's own help documentation admits it: the company "uses your history to provide, develop, and improve its services (such as training generative AI models)." The email to customers put it in plainer language: "Like your Search Services History, your saved media is also used to develop and improve Google services and technologies, including AI models and safety measures."
Safety measures. A convenient catch-all.
Google isn't alone. Meta trains its models on public Instagram and Facebook posts — and on images captured by its Ray-Ban smart glasses. The pattern is clear: consumer tech giants have exhausted the open web. They've scraped Wikipedia, Reddit, Common Crawl, the entirety of digitized human knowledge. Now they need fresh, proprietary, high-signal data. And they own the pipelines that generate it.
Your phone is the sensor. Their apps are the collectors. The terms of service are the permission slip you never read.
This shift — from scraping to harvesting — changes the power dynamic. Web scraping is parasitic but distant. Harvesting user uploads is intimate. It requires trust. Google is spending that trust like a drunken sailor.
You can stop it. Google will tell you that. The controls exist. But they're scattered across Search Services History and Search Services Personalization pages — separate URLs, separate toggles, separate mental models. Uncheck "Save Media." Uncheck "Search Services History." Set auto-delete to three months. Do it on every device. Do it for every account.
Friction is a feature. Every extra click, every separate page, every confusing label reduces the opt-out rate. Google knows this. It designs for this.
And even if you disable everything today, the data already saved remains. Google's language: "Some of this storage is temporary and tied to making the product work, but per Google's own language, saved media can also be retained specifically to train its AI." The past is not yours to recall.
The June email framed the change as giving users "more control." That's the tell. When a company takes more and calls it control, it's lying. Real control means opt-in. Real control means a single toggle labeled "Use my media to train AI" — off by default. Real control means deleting the data already collected.
Google offers none of those. It offers a maze of checkboxes and a hope you won't bother.
Regulators are watching. The EU's GDPR requires freely given, specific, informed consent for processing personal data. The UK's ICO has already warned companies that legitimate interest won't cover AI training on user content. California's CCPA gives residents the right to opt out of sale or sharing. But enforcement lags. And Google's lawyers are excellent at threading needles.
We've accepted a bargain: free services in exchange for advertising profiles. That bargain is mutating. Now the payment is your creative output, your voice, your visual world — fed into models that will eventually compete with you, replace you, or surveil you.
Google's AI Overviews already summarize the web so you don't visit it. Its generated images already displace photographers. Its code assistants already undercut developers. The media you upload today improves the systems that devalue your labor tomorrow.
Opting out is necessary. It's not sufficient. The only durable fix is structural: laws that treat AI training on personal data as what it is — a use that demands affirmative consent, not a pre-checked box in a settings menu you'll never open.
Until then, go to your Google Account. Search "Search Services History." Turn it off. Turn off "Save Media." Set auto-delete to three months. Do it now.
Then ask yourself why you had to.