Google is rolling out new privacy settings for Search services that will change how the company handles user history data.
The email sent to users recently carries the subject line “New privacy settings for Search services.” This rollout is happening globally over the next few months. It alters the rules for how Google retains and uses Search history information.
Every piece of media uploaded to your account, including photos for reverse image searches and audio recordings from Google Translate, may now be kept and fed into Google’s AI models.
A new option called Search Services History was already enabled when the author first visited the settings page. The box to save all uploaded media for AI training was also checked by default. This is the standard state for most users.
To change this, you must visit the My Activity page and select the Search Services History tab. This area shows what Google saves from your history. It is where you can turn off the entire setting or delete your activity. It is critical to uncheck the box next to Save media if you do not want your image uploads used for training.
It is worth making this change now rather than waiting. Once your media data is processed, it is difficult to reverse. A pop-up explains that if saved media is used to train models, it is disconnected from your Google Account. The training data will be kept for up to four years, even if you delete the original activity. That is a long time for random image uploads to remain in the digital ether.
Davis Thompson, a Google spokesperson, said in an email that these settings help users get more relevant results and revisit searches, including visual and voice ones. He noted they can be turned on or off at any time. He did not answer a question about why the feature is on by default.
Google states it is starting to store much more than just typed text. Your saved media includes images, files, and audio or video recordings from interactions with Search services. This covers Google Lens images, recordings from Search Live or Translate speaking practice, content you upload, and voice searches.
AI models need diverse inputs beyond text, such as audio or video. If Google gathers more data and more types of data from its users, it may innovate faster than its competitors.
Google’s massive user base across multiple services gives it an edge in data collection. Thorin Klosowski, a senior security and privacy activist at the Electronic Frontier Foundation, noted that Google is in a unique spot compared to other companies. People have been using these services for a long time and have grown comfortable with the amount of data collected. Apps that people use every day have built-in inertia, so changes users dislike may not drive them to alternative services.
Being forced to opt out of AI training seems to be the standard across sites and platforms. Klosowski argues that asking users to consciously choose to enable these features is the least companies can do. Google would have to make a stronger case for why these features are helpful if they were not automatically turned on.
In an email sent to a testing account on June 23, the first sentence framed the change as giving the user “even more control over saved history.” The message provided examples of how saving media may be helpful, such as revisiting past visual searches with Lens or continuing a Search Live conversation about a song. Google did not provide similar examples after stating near the end of the email that the saved media will be used for AI model training. The message simply continued to the next detail.
This is another major software change worth slowing down to process for everyday users. Ben Winters, director of AI and privacy at the Consumer Federation of America, said it creates an extra layer of math a consumer has to do about whether they feel comfortable using the tool they have used for a long time.
The author feels constantly overwhelmed that they must opt out of data training for every service. It leaves them feeling like they are likely to miss something buried in all these settings.
Winters sees this change as placing the onus on users to avoid AI training. This may contribute to widespread user exhaustion. “There’s an increasing feeling of powerlessness and hopelessness about even trying to protect your data, because every little thing is going to be squeezed out of you,” he said.
What it means
Users must actively manage their privacy settings to prevent personal media from being used to train AI models. The default setting now allows Google to keep this data for up to four years after deletion from the account. This shift moves the responsibility of data protection entirely onto the individual.
