Chrome Quietly Drops Privacy Promise for On-Device AI – What You Need to Know

If you use Google Chrome, you might have noticed a new section in settings labeled “AI.” Over the past few months, Google has been rolling out on-device AI features like “Help me write” and “Tab compare.” The company originally stated clearly that these features process everything locally, without sending data to Google’s servers. That explicit promise has now disappeared, and several news outlets reported the change in early May 2026.

Here’s what happened, why it matters for your privacy, and how you can adjust your browser settings to stay in control.

What Happened

On May 7, 2026, Decrypt reported that Google had silently removed a key sentence from Chrome’s privacy FAQ. The deleted line read: “All processing happens locally on your device, and no data is sent to Google servers.” This statement had been present since the on-device AI features were introduced. The removal was later confirmed by the Japanese tech site GIGAZINE on May 9, 2026.

The timing is notable. Just a day earlier, on May 6, Decrypt also reported that Chrome had been quietly installing a 4GB AI model on users’ computers—and that even after deletion, the browser would reinstall it. Google admitted that the model “may still communicate with Google servers” for tasks like model updates or telemetry, but originally framed this as separate from the actual AI processing.

By removing the blanket privacy promise, Google no longer guarantees that your interactions with these AI features stay 100% offline. The company has not issued a public statement explaining why the line was removed, leaving users to draw their own conclusions.

Why It Matters

The phrase “on-device AI” is often used to reassure users that their data remains private. When a browser like Chrome promises local processing, it implies that even Google cannot see what you type or what suggestions it generates. Removing that promise doesn’t automatically mean data is now being sent to servers—but it removes a layer of transparency and trust.

Many users may not have realized that the AI features were enabled by default. Chrome began automatically installing the 4GB model in background updates. Even if you never clicked “Enable,” the browser could have downloaded the model and, potentially, activated certain AI functions. Combined with the privacy FAQ change, the situation raises legitimate questions:

  • Are my inputs being logged or analyzed remotely?
  • Does Google use this data to train its larger models?
  • Can I trust the “on-device” label going forward?

Google has not provided answers to these specific questions. Until it does, the safest approach is to assume that your data may be at risk—especially if you use features that rely on the cloud-connected model.

What You Can Do

You can take control of these settings right now. Chrome provides a dedicated AI settings page that lists all installed models and lets you disable individual features or the entire on-device AI system.

Step 1: Open the AI settings page
Type chrome://settings/ai into your address bar and press Enter. On some versions, you can also go to Settings > AI from the main menu.

Step 2: Review the available options
You’ll see features like “Help me write,” “Tab compare,” and “Organize tabs.” Each has a toggle. The page also shows whether the 4GB on-device model is installed.

Step 3: Disable what you don’t want
Turn off any feature you don’t actively use. If you want to be cautious, disable all of them. Chrome will still work normally; these are entirely optional.

Step 4: Check for the installed model
If you see the AI model listed, you can try deleting it from the same page. However, based on reports, Chrome may re-download it on restart. A more reliable method is to use the browser’s built-in “Reset and clean up” option, but that also resets other settings. For now, the best defense is to keep the features turned off.

Step 5: Monitor future updates
Google may add new AI features in future Chrome releases. Periodically revisit chrome://settings/ai to see what’s new. Also consider using privacy-focused browsers like Brave or Firefox if you want stronger default protections.

Broader Context

This isn’t the first time Google has been criticized for shifting privacy boundaries. The company has a long history of adding features that collect data, then adjusting privacy policies later. The deletion of the on-device AI promise fits that pattern.

Other browsers have taken a different approach. For example, Apple’s Safari and Microsoft Edge have both introduced on-device AI features but have maintained clear statements about local processing. Google’s move stands out because of the silent nature of the change and the automatic download of a large model.

Until Google clarifies its new privacy stance, users should treat Chrome’s AI features with caution. The tools themselves can be useful—but not at the cost of your privacy.

Sources

  • Decrypt: “Chrome Deleted Its Own Privacy Promise for Sneaky On-Device AI” (May 7, 2026)
  • Decrypt: “Chrome Is Quietly Installing a 4GB AI Model on Your Computer—And Putting It Back If You Delete It” (May 6, 2026)
  • GIGAZINE: “Chrome removes claim that its ‘on-device AI’ does not send data to Google servers” (May 9, 2026)