Google uses user-uploaded search files to train AI

Google has introduced changes that allow it to use media files uploaded by users through its search services to train artificial intelligence models. 

Engadget reports.

The new feature is enabled for all users by default.

What data the company collects

According to Engadget, the update covers a wide range of content that users interact with across the company’s search services. This list includes images, files, as well as audio and video recordings.

The publication notes that:

• if a user uploads a photo to Google Lens for visual search, the company may use it to train models;
• a similar approach applies to audio recordings that accompany Google voice search;
• any files uploaded to the Google Translate service are also subject to collection.

These changes apply exclusively to products directly related to the Search system. Users’ personal files stored in Google Photos are not subject to this data collection.

How to disable data collection

Because the media collection feature is enabled automatically for every profile, users who wish to opt out must change the settings themselves.

To disable data collection, you need to:

1. Go to the special Search Services History page and uncheck the «Save Media» option.
2. Go to the Search Services Personalization page and ensure that the system is not saving any information.

In addition, users can disable the display of results from the AI Overview features.

Remember that Cloudflare will change its policy on webcrawlers that both index sites for search engines and are used to train or run artificial intelligence models.

The hosting platform will block such webcrawlers starting in September this year. The company aims to return control over the use of site content to site owners.

Also read: The Ministry of Digital Transformation is preparing an AI law. How is AI currently regulated in Ukraine?