Chrome Quietly Removed Its Privacy Promise for On-Device AI: What You Need to Know
Google Chrome recently updated its privacy documentation for on-device AI features, and one key promise quietly disappeared. The original language stated that on-device AI features “do not send your data to Google servers.” That guarantee has been replaced with broader wording that allows data use for “personalization.” If you use Chrome and care about how your browsing data is handled, this change is worth knowing about.
What happened
On May 7, 2026, Decrypt reported that Google had altered Chrome’s privacy description for on-device AI. The previous version explicitly assured users that no data leaves their device. The revised version removes that assurance and instead describes how data may be used to improve AI features, including personalization.
The exact new phrasing has not been confirmed by Google at the time of writing, but screenshots and summaries from several news outlets (including Decrypt and Yahoo Tech) indicate the commitment to “no data sent to servers” is no longer there. Google has not issued a public statement about the change, so the motivation is unclear. It may be related to planned AI features that rely on some server-side processing, or it could be a legal clarification.
Why it matters
On-device AI is often marketed as a privacy-friendly approach: the processing happens locally, so your data stays with you. The previous promise was a simple, reassuring line that helped users trust Chrome’s AI tools.
Now that the promise is gone, the question becomes: what exactly changes in practice? If Chrome’s on-device AI features start sending data to Google’s servers—even in aggregated or anonymized form—that would be a meaningful shift. For users who rely on Chrome’s default settings, the change could mean their browsing habits, search queries, or other local data are used for purposes they didn’t explicitly consent to.
This also matters because privacy promises are supposed to be stable. When a company quietly revises a core commitment without fanfare, it erodes trust—especially for something as sensitive as browser data.
What readers can do
You don’t have to accept the new terms without review. Here are concrete steps you can take to limit data collection in Chrome:
Check your on-device AI settings. Go to
chrome://settings/ai(or navigate through Settings → Advanced → AI features). Look for any toggles related to on-device AI. If you see options to disable “help improve AI features” or “personalization,” turn them off.Review your privacy and security settings. In
chrome://settings/privacy, check the “Privacy Sandbox” and “Cookies and other site data” sections. Disable any ad personalization or tracking features you don’t want.Consider alternative browsers. If Chrome’s direction concerns you, browsers like Firefox, Brave, or Vivaldi offer similar AI features with more explicit privacy controls or no server-side data use. For now, the change is a documentation update, but it may foreshadow features that require sending data.
Keep an eye on official announcements. Google may clarify the change in the coming weeks. Before making any drastic switches, wait for a clear statement about what data is actually collected and how it’s used.
Broader context
This is not an isolated incident. As AI features become more common inside browsers, companies are reevaluating old privacy promises. Microsoft, Apple, and Mozilla have all updated their AI language in recent years. The challenge is balancing powerful AI features (like smart tab grouping, writing assistance, or search summarization) with the promise of local-only processing.
For now, the Chrome change is mostly symbolic: the existing on-device AI features likely haven’t changed behavior yet. But the removal of a clear guarantee should make users pause. When a company deletes a simple, reassuring statement, it’s reasonable to ask what they plan to do next.
Sources: Decrypt (May 7, 2026); Yahoo Tech (May 7, 2026). Google has not publicly commented on the change as of this writing.