2 Sources
[1]
Google is training AI on even more of your data now, unless you opt out - here's how
Follow ZDNET: Add us as a preferred source on Google. ZDNET's key takeaways * Google has again changed its policy regarding how it uses your data. * Images, voice searches, and videos can be used to train LLMs * You can disable this feature to retain your privacy. When you upload any media to Google's search services, it's now being used to train AI. The result? Less privacy -- unless you opt out. Google quietly updated its search privacy settings last month, as announced in an email, and automatically opted all its customers into its expanded AI training. Here's the updated documentation: Also: 13 Google Photos settings I always change on every new device - and why "You can choose whether Google saves media to your Search Services History when you're signed in. Saved media includes your images, files, audio, and video from your interactions with Search services. Your media may be used to improve your experience on Google services, like letting you revisit your past visual searches. Saved media may be used to develop and improve Google's AI models and technologies, as well as the Google services that use them. When media is saved, you can view it in your Search Services History. Learn how to find and manage Search Services History." What does this mean? This new language essentially means that any media generated by your search using Google's services can be used for any purpose, including future searches and training AI. When you combine that with the current targeted advertising, a picture emerges of less and less privacy. Also: Google's new AI Search box is here - along with agents and 5 more upgrades Let's say you're discussing with a chatbot the idea of leaving your job. That discussion is used to further train the chatbot's LLMs. Suppose you share specific details, such as the name of your company, your boss's identity, and why you don't like working there. The LLM is being trained on all that information. A few days later, your boss happens to ask the same LLM -- the one trained with your previous chat -- a question about employee loyalty. In the chatbot's response, your boss sees some rather familiar information that could perhaps make him put two and two together... All of a sudden, your plans are doxxed. Your privacy is compromised. Google's update is not only about the media that appears in response to your searches. It also applies to media you upload for AI searches. For example, you upload screenshots taken with Circle to Search, photos you've taken, or use voice searches (via Search Live or using Google Translate to practice learning a new language); even that data is now saved to train LLMs. Consider this: Your voice is used to train LLMs. Imagine if someone were to decide to use AI to create an audio clip of you saying something you didn't say. Also: Sick of AI in Search? These 7 Google alternatives still put links first Dystopian? Yes. Reality? Also, possibly yes. All this is being done in to improve Google's AI searches. What can you do? The most important thing you can do is opt out of this feature/service. Doing so will ensure that those media files are not used to train LLMs. Also: How to remove AI Overviews from Google Search: 4 easy ways There are two pages within Settings to focus on: Search Service History and Search Service Personalization. For the Search Service History page, open the Google app, tap your profile image in the top right corner, then tap "Search history." On that page, make sure "Save media" is unchecked, then tap the top entry for Search Service History. If you see that it's listed as On, tap that button, and then you can turn the service off. For Search Service Personalization, go back to the Google app, tap your profile image, and then tap "Search personalization." On this page, tap the top entry and, from the resulting page, make sure the service is turned off. If you can't find these pages (or if your device doesn't allow you to disable the services here), there's another option. Head over to myactivity.google.com, click Search Services History, and uncheck the box for Save media. I keep all of those options disabled to ensure my privacy. If you decide to leave any of these features on, you can configure the duration for which that information is retained: 3 months, 18 months, or 36 months. The good news is you have some control here. You can change your preferences on the Search Services History and Search Services Personalization pages. On the former, you can uncheck the "Save media" box separately from the Search Services History box, or uncheck both. You can also configure how often you want saved data automatically deleted -- after 3, 18, or 36 months. Finally, go to myadcenter.google.com and ensure the Personalized Ads feature is disabled. Are there downsides to this? If you consider Google's ability to predict what you might want to see in advertisements a positive thing, or if you like the idea of the company using your images, voice, videos, and other files to train its large language models, then you might consider leaving these features enabled. Likewise, if you like personalized ads (because they make shopping easier), leave those settings enabled. Also: 5 reasons I use local AI on my desktop - instead of ChatGPT, Gemini, or Claude This is a half-full/half-empty scenario that depends on what level of privacy you are comfortable with. For me, it's very much a half-empty situation, but that is because I consider my privacy to be of utmost importance. The need for as much privacy as I can muster is also part of why I only use locally installed AI. If you consider privacy as important as I do, consider disabling this feature so Google's LLMs don't train on your saved media. If you really want to up your privacy, consider only using locally installed AI.
[2]
Google's AI Has Access to More Than You Think. Change These 7 Settings Now to Protect Your Privacy
Do you ever use Google Gemini? If so, you've probably noticed that the AI chatbot is eager to learn as much as possible about you to better personalize its responses. But how much does this AI chatbot really need to know? First, you should know what information Google's AI collects. This list includes prompts, shared files, videos, photos, browser pages, transcripts, Gemini Live recordings, the names and custom instructions of your Gems, and any additional instructions you give Gemini to tailor its responses. Keep in mind that Gemini also powers many of Google's base features now, so even if you don't use the chatbot, you're likely interacting with the AI model in some way. Google offers a few ways to control your privacy, but it does make certain changes difficult. For example, you can delete past conversations and revoke Gemini's access to Google Drive, Gmail, and Google Calendar. You can also use Gemini in temporary mode so it won't save your chats. However, the only way to prevent Gemini from training on your chats is to turn off Google's history retention for them. And if you do that, it means you won't be able to view or refer back to prior conversations. That said, there are some workarounds that strike a balance between privacy and effectiveness when using Gemini. Here are seven ways to boost your privacy when using Google's AI platform. 1. Opt Out of Google Search AI Training In its quest to gather even more of your information, Google is now using your search uploads to train its AI models. That means every time you add an image, document, audio file, or video to the Google search bar, it is collected in order to improve Gemini-powered products like Google Lens and Search Live. This is now happening by default, so if you don't want Google's AI to know what you're looking for online, you'll need to disable the feature. To do this, open the My Activity settings screen, choose Search Services History, and make sure the Save media option is unchecked. You can also turn Search Services off entirely, delete past activity, and set up an auto-delete option. 2. Don't Let Google Use Your Chats for Training Gemini will use your conversations with it to train itself. Though Google promises that your data is anonymized, you may still not want any sensitive or personal details being shared this way. Unfortunately, the only way to stop this is to completely prevent Google from storing and using previous chats. This will remove conversations from Google's activity history and your own chat history, meaning you won't see any of your previous chats. To me, that's a huge drawback as I often refer to my chat history to see what topics have already been covered. But if you don't need to view your chat history, sign in to the Gemini Apps Activity page with your Google account. If the Keep activity setting is on, click the drop-down button and choose Turn off. The next window explains what happens when you turn off your Gemini Activity. Beyond preventing future conversations from being used for AI training, this action removes these chats from your Activity feed and chat history. To confirm this step, click the Got it button. When you return to the main Gemini screen, you'll see that your previous chats are gone. 3. Delete Past Chats From Google's History Since I want to access many of my previous chats, I typically don't disable them from Google's activity. Instead, I'll selectively remove certain conversations that I no longer need. Google lets you delete specific chats, all those from a certain period, or everything in one shot. Deleting chats this way removes them from Google's activity history and your chat history. But at least it gives you control over which conversations you want to remove and which ones you wish to keep. At the Gemini Apps Activity page, I'll remove certain conversations that I feel are more sensitive or personal and don't want Google to store, either for its own activity or in my history. Click the X next to any individual chat you wish to delete. To remove all chats or just those from a specific time period, click the Delete drop-down button near the top of the Activity page. You can then opt to delete activity from the last hour, the last day, all time, or a custom range. Instead of manually removing conversations, you can also set up an auto-delete option. Near the top of the Activity page is a setting called Choose an auto-delete option if the option is currently turned off or Deleting activity older than... followed by a specific period of time if the option is currently on. Click the right arrow next to that setting and choose Auto-delete activity older than from the drop-down menu, and choose a specific timeframe -- 3 months, 18 months, or 36 months. Click Next > Confirm > Done. If you don't want your chat history to be removed automatically, choose Don't auto-delete activity instead. 4. Prevent Gemini Live From Storing Your Voice Conversations Gemini Live allows you to converse with Google's AI by voice. But here's another area where the company can use your chats for AI training. By default, Google captures audio recordings of your conversations and any screenshots shared during a live conversation. Some recordings are even sent (anonymously) to human reviewers. Thankfully, you can easily turn this option off. At the Gemini Apps Activity page, look at the top of the screen for a setting called Improve Google services with your audio and Gemini Live recordings. If it's checked, click the box to uncheck it. Read the information displayed at the pop-up window. If this sounds good to you, click the Turn off button to finalize your action. 5. Cut Off Gemini's Access to Google Drive, Gmail, and Google Calendar Gemini can access your files and other content from certain connected apps, including Google Drive, Gmail, and Google Calendar. Part of Gemini's Personal Intelligence skill, this lets you run AI-related tasks and actions on your own data. For example, you could tell Gemini to summarize a particular file stored on Google Drive or to generate a response to a certain email in Gmail. If you don't want Gemini to have access to your files, you can easily disable this. At the main screen, click Settings & help > Personal Intelligence > Connected Apps, and then turn off access to Google Workspace (Gmail, Google Docs, Google Keep, Google Calendar, Google Drive, and Google Tasks), Google Photos, Google Search, YouTube, and YouTube Music. Turn off the switch for any apps and services that you don't need to use with Gemini, and then select Disconnect. 6. Turn Off Personal Intelligence's Memory Feature Another aspect of Gemini's Personal Intelligence is Memory. With this enabled, Gemini remembers your past chats and learns from them to better understand your interests, preferences, and other personal details. The goal here is to help the AI tailor its responses to you, but you might not want this. This is an opt-in feature, so it should already be disabled. To check, click Settings & help > Personal Intelligence, then check whether the Memory switch is enabled. If it is, you can just turn it off. 7. Use Gemini in Temporary Mode One more way to protect your privacy with Gemini is to use it in temporary mode. You can easily start a new temporary chat to discuss a private or sensitive topic; it won't appear in your chat history or on the Google Activity page, and won't be used to train Gemini or personalize your experience. You can set one up by clicking the Temporary chat button next to New chat in the left sidebar. Gemini explains how the temporary chat works. Type and submit your request, knowing that the chat will remain private. Temporary chats are retained only for 72 hours.
Share
Copy Link
Google quietly updated its privacy policy last month, automatically opting users into expanded AI training that includes images, audio files, and videos from search interactions. The change affects Google's Gemini-powered services and raises significant privacy concerns. Users can disable these features through Search Service History and Gemini Apps Activity settings, though doing so requires manual action across multiple pages.
Google AI has quietly expanded its data collection practices, now using images, audio files, videos, and other user-uploaded media to train its large language models. The policy change, announced via email last month, automatically opted all users into the expanded AI training program
1
. According to updated documentation, saved media from search interactions can now be used to "develop and improve Google's AI models and technologies, as well as the Google services that use them"1
.This expansion affects various Google services, including screenshots from Circle to Search, voice recordings from Search Live, Google Translate practice sessions, and photos uploaded for visual searches. The change represents a significant shift in Google's privacy policy, raising data privacy concerns about how personal information flows into AI systems
1
.
Source: PC Magazine
The privacy concerns extend beyond simple data collection. When users interact with Gemini, Google's AI chatbot, the system collects prompts, shared files, browser pages, transcripts, voice recordings, and custom instructions
2
. This information trains AI models that could potentially expose sensitive details. Consider a scenario where someone discusses leaving their job with specific company details and boss information. That conversation trains the AI model, which could inadvertently surface similar information when another user asks related questions, creating potential privacy breaches1
.Voice recordings present another dimension of privacy risks. Your voice can be used to train AI systems, raising concerns about potential misuse for creating synthetic audio clips
1
. Even users who don't directly interact with Gemini are affected, as the AI model now powers many of Google's base features2
.Users can disable features through multiple settings pages to regain control over their data. For Search Service History, open the Google app, tap your profile image, select "Search history," and ensure "Save media" is unchecked
1
. Alternatively, visit myactivity.google.com, click Search Services History, and uncheck the Save media box1
.
Source: ZDNet
For Gemini-specific data collection practices, users must access the Gemini Apps Activity page. The only way to prevent Gemini from using conversations for AI training is to turn off the Keep activity setting entirely
2
. This action removes chats from Google's activity history and your chat history, meaning you won't be able to view previous conversations—a significant tradeoff for privacy-conscious users who reference past interactions2
.Users can also selectively delete specific conversations or set auto-delete options for 3 months, 18 months, or 36 months
1
2
. Additionally, disabling personalized ads at myadcenter.google.com provides another layer of privacy protection1
.Related Stories
The policy shift highlights a broader trend where tech companies balance AI improvement against user privacy. While Google promises data anonymization, the automatic opt-in approach places the burden on users to actively protect their information. Short-term, users face immediate decisions about whether to sacrifice chat history access for privacy. Long-term, these data collection practices could influence how comfortable people feel using AI assistants for sensitive topics like health concerns, financial planning, or workplace issues. Watch for potential regulatory responses as privacy advocates scrutinize these expanded data collection practices and whether other AI providers follow similar paths.
Summarized by
Navi
06 Jul 2026•Technology

16 Jun 2026•Technology

14 Aug 2025•Technology

1
Technology

2
Business and Economy

3
Technology
