Google will save your Lens photos, Search Live recordings, and Translate audio for AI training
Google is expanding its data collection infrastructure by automatically retaining user interactions across Lens, Search Live, and Translate services under a new 'Search Services History' setting. This move signals Google's commitment to building proprietary training datasets from consumer touchpoints, directly competing with rivals who rely on web-scraped or licensed data. The policy shift raises questions about consent granularity and whether users understand the dual purpose of their searches as both service logs and model training material. For AI practitioners, this represents a major player consolidating first-party data advantages at scale.
Modelwire context
Analyst takeThe detail worth sitting with is that Search Services History bundles three distinct modalities (visual, audio, and text) into a single retention setting, which means users who want to opt out of one must opt out of all. That bundling is a consent design choice, not a technical necessity, and it substantially raises the floor on how much data Google collects by default.
This is largely disconnected from recent activity in our archive, as we have no prior coverage to anchor it to. It belongs, however, to a broader pattern playing out across the industry: the race to secure non-scraped, interaction-derived training data as web-crawl pipelines face legal pressure and quality ceilings. Google's move here is structurally similar to what Meta has done with Ray-Ban footage and what Apple is doing with on-device query logs, though the scale of Google's consumer surface makes the data volume incomparable.
Watch whether the EU's data protection authorities issue a formal inquiry within the next six months, specifically targeting the bundled opt-out design. A challenge there would force Google to disaggregate the setting and would set a precedent that constrains how every other major platform structures multimodal data retention.
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsGoogle · Google Lens · Google Search Live · Google Translate · Search Services History
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. The full content lives on theverge.com. If you’re a publisher and want a different summarization policy for your work, see our takedown page.