How to Stop Google from Using Your Photos to Train AI

by priyanka.patel tech editor

For millions of users, Google Photos is more than a digital shoebox; it is an intelligent archive that can instantly surface a “mountain trip” or a “birthday party” from a decade ago. But that convenience comes with a quiet trade-off. If you utilize the service, there is a significant chance your personal images are being used to train Google’s artificial intelligence—often through default settings that users never explicitly opted into.

The core of the issue lies in the intersection of cloud storage and generative AI. Although Google maintains that photos stored simply within the Google Photos library are not used to train AI models operating outside that specific app, the boundary blurs the moment you interact with Google’s broader AI ecosystem. Specifically, since late 2025, Google has used photos, files, and prompts shared through the Gemini app to refine and improve its AI models.

In other words that if you upload a photo from your library into Gemini to ask a question or link your Google apps for a more seamless experience, that image is no longer confined to your private library. It enters a different set of terms where Google can utilize the data for training. For those concerned about digital privacy, the good news is that you can stop Google from using your photos to train AI, though the controls are scattered across several different menus.

The Gemini app allows users to interact with their Google Photos library, but this interaction can trigger AI training protocols. Credit: Digvijay Kumar / MakeUseOf

Closing the Gemini Data Loop

The most direct way to prevent your media and prompts from being used for AI training is through the Gemini settings. In September 2025, Google rebranded the “Gemini Apps Activity” setting to “Keep Activity.” This change may have confused users who were looking for the previous terminology, but it remains the primary toggle for data harvesting.

Closing the Gemini Data Loop

By default, Google uses a sample of conversations and uploaded media to improve its services. To disable this, users should visit gemini.google.com, click the hamburger icon (three horizontal lines) in the top left, and navigate to Settings & support and then Activity. From the dropdown menu, you can select Turn off or Turn off and delete activity to remove previously saved data.

even after disabling this setting, Google retains recent conversations for up to 72 hours for operational reasons before they are permanently deleted. For those who want a completely clean slate for a specific interaction, Google offers “Temporary Chats,” which function similarly to an incognito session and do not save data to your history.

Managing AI Features Within Google Photos

Beyond the Gemini app, there is a specific set of AI-driven tools embedded directly within the Google Photos app. One of the most prominent is “Ask Photos,” a conversational tool that allows users to search their library using natural language (e.g., “Where did we stay on our last vacation?”).

While Google asserts that personal data stored within Google Photos is not used to train AI models that operate outside of the app, these features are enabled by default and grant the system deep access to your library to generate narrated recap videos and personalized memories. To review these permissions, open Google Photos, tap the profile icon, go to Settings, select Preferences, and then tap Gemini features in Photos.

Within this menu, users can disable “Ask Photos” and “Gemini-powered memories” entirely. There is also a specific toggle titled Allow us to access your queries. When active, Google may review the search queries you enter to improve the feature; while these queries are not tied to your account or inclusive of the actual photos, disabling this toggle provides an extra layer of privacy.

The Broader Workspace Privacy Risk

Privacy concerns aren’t limited to imagery. Google employs “Smart Features” across its Workspace suite—including Gmail and Google Drive—to power AI-assisted tools. These features allow Gmail to suggest sentence completions, automatically detect travel bookings for your calendar, and offer customized responses based on your historical email behavior. In Google Drive, this manifests as automated document summaries and smart suggestions.

As these are enabled by default, Google’s systems essentially process your emails and documents to create these tools functional. To opt out, users can open Google Drive, click the gear icon, navigate to Settings, then Privacy, and click Manage Workspace smart feature settings. Turning off both “Smart features in Google Workspace” and “Smart features in other Google products” will stop these AI-assisted tools from functioning.

Summary of Privacy Toggles

Key Google AI Privacy Controls
Feature Area Setting Name Primary Impact
Gemini App Keep Activity Stops training on prompts and uploaded media.
Google Photos Gemini features in Photos Disables AI-driven library searches and recaps.
Workspace Smart features Prevents AI processing of emails and documents.
Account Wide Privacy Checkup Centralized review of major data settings.

Taking Full Control of Your Digital Footprint

Because Google distributes these settings across different apps and menus, a comprehensive privacy audit is often the only way to ensure total opt-out. For a broader cleanup, users can visit myactivity.google.com to review and delete saved activity across all Google products. This page also allows for the setup of an auto-delete timer, which can automatically purge data older than three or 18 months.

For those overwhelmed by individual menus, the Privacy Checkup tool at myaccount.google.com provides a guided walkthrough of the most critical data settings. As a former software engineer, I have seen how “dark patterns”—design choices that nudge users toward less private options—are used to keep default settings active. The burden of privacy remains with the user, but the tools to reclaim it are available.

As Google continues to integrate Gemini more deeply into the Android OS and Workspace, the boundaries between “storage” and “training data” will likely continue to shift. Users should keep an eye on updated Google Privacy Policy terms, as changes to how generative AI handles personal data are frequently updated to comply with evolving laws in the EU and UK.

We want to hear from you: Have you noticed your Google Photos becoming more “predictive” lately? Let us understand in the comments if you’ve adjusted these settings or if you find the process too cumbersome.

You may also like

Leave a Comment