How Google’s AI training uses your Search interactions — and how to opt out

Google has quietly expanded the types of interactions it logs for AI training. Beyond traditional search queries, the company now captures images from Lens searches, audio from real-time voice queries, and phrases spoken into its Translate app — all stored under a new Search Services History setting. While the data helps improve features like voice recognition and image search, users can disable the feature entirely.

Why Google is saving more Search data for AI

The expansion reflects Google’s broader push to use real-world inputs to train its AI systems. Images scanned via Google Lens, audio clips from Search Live recordings, and phrases spoken into the Translate app are now included in the data pipeline. Previously, such interactions might have been stored separately or not at all. Google states these records are used to "provide, maintain, and improve" its services, including refining voice-to-text accuracy and image recognition models.

Key changes include:

• Google Lens images: Photos you upload or screenshot for visual searches are now logged.
• Search Live recordings: Real-time voice queries are stored alongside other audio inputs.
• Translate phrases: Spoken phrases entered into Google Translate are captured for model training.

How to review and disable Search Services History

Users who wish to stop Google from saving these interactions can adjust their settings in minutes. Google provides a dedicated toggle under Search Services History, with additional granular controls for media-specific storage. For example, disabling the "Save Media" option prevents images and audio from being retained. These changes apply retroactively, meaning existing data can be cleared manually if needed.

To toggle off Search Services History:

1. Open Google Account settings.
2. Navigate to Data & Privacy > History Settings > Search Services History.
3. Turn off the toggle or disable "Save Media" for specific media types.
4. Optionally, delete past interactions from the same menu.

Privacy trade-offs: AI benefits vs. user control

Google’s move highlights the tension between AI advancement and privacy expectations. While aggregated and anonymized data fuels improvements in features like voice search and Lens accuracy, critics argue transparency and user agency remain inconsistent. The company insists data is handled in compliance with its privacy policies, but advocates call for clearer opt-out mechanisms.

For now, users must proactively adjust settings to limit data collection. As AI integration deepens across Google’s ecosystem, expect further expansions in what interactions are logged. Those uncomfortable with passive data sharing may find relief in periodic reviews of their Search Services History and other Google account activity.

Looking ahead, the balance between innovation and privacy will likely shape future policy decisions, with regulators and users both weighing in on acceptable boundaries for AI training data.