Google has expanded the scope of data it collects to train its AI models, now including media uploaded by users across many of its core search-related services.
the Policy changeIt was implemented without much public fanfare and allows the tech giant to use photos, audio, video and other files submitted through tools like Google Lens and Google Translate, Engadget reported on Monday (July 6).
Google’s move highlights the demand for high-quality datasets as creative AI developers face a dearth of new information to feed their large language models.
Under the updated terms, any image uploaded to Google Lens for visual identification or audio captured during voice search may be collected for training purposes. Data collection also extends to any files processed through Google Translate, including “images, files, audio and video recordings,” according to the report.
For professionals in the digital economy and banking sectors concerned about data privacy or corporate security, it is worth noting that users are automatically selected for this training program. Engadgetciting his previous findings TechCrunchindicates that the current policy is limited to search-related products; Personal repositories such as Google Images are currently excluded from this specific training data sweep.
As generative AI searches for new data sources, Google has provided a manual mechanism for users to restrict the use of their data in this way. To opt out, users must go to their designated Search Services history page to uncheck the “Save Media” box. Additionally, users are advised to review their search service personalization settings to ensure that no other AI training media is retained.
For those seeking to limit their interaction with Google’s AI outputs entirely, the report also highlights a technical solution: appending “-AI” to a search query will effectively remove the AI-generated overview results from the interface.
This shift highlights a broader trend among big tech companies seeking to leverage proprietary user interactions to maintain a competitive edge in the AI race, even as questions surrounding user permission and data ownership persist. Google itself highlighted this trend earlier this year, when the company compressed News organizations to let their AI train on their articles or risk losing the annual payment to appear in Google News.
