Google Drive Just Got a Massive AI-Powered Scanner Upgrade — Here’s Everything You Need to Know

Google has just completed a fundamental rethinking of document scanning on Android. After months of public beta testing and incremental user interface tweaks, a powerful new AI-driven scanner is now widely rolling out inside the Google Drive and Files by Google apps. Instead of tapping a shutter button for each page, you can now simply flip through a stack of documents in front of your phone’s camera. Three intelligent features handle the capture, quality control, and organization automatically — and they do it without sending a single image to the cloud. The catch is that these smarts need serious on-device horsepower, which means they are reserved for newer, higher-spec Android phones .

Three new AI-powered features

The scanner’s intelligence is built around a trio of capabilities designed to remove the friction from digitizing long documents .

Smart Batch Scanning is the headline change. You no longer need to manually frame and capture every single page. Once the scanner is active, hover the camera over your document and the system will automatically detect when a new page is placed in view, capture it, and prepare for the next one. The experience feels more like recording a video than taking a series of photographs .

Auto-Best Frame works quietly in the background during a batch capture. If you move, the lighting changes, or a page is momentarily blurry, the scanner automatically selects a clearer frame from the same capture session. You won’t need to reshoot a smudged or poorly focused document, because the scanner already has a better version ready .

Duplicate Detection acts as a safety net at the end of the scanning process. If you accidentally passed the same page in front of the camera twice, the feature identifies the near-duplicate and asks whether you want to keep it or remove it before saving the final PDF. This prevents the frustration of discovering a doubled page after you’ve already filed the document .

These three features work together to move the scanning experience from a manual, start-and-stop chore to a largely automated flow.

Full on-device processing for speed and privacy

A key differentiator for this update is that none of the new AI processing relies on an internet connection. All three features — batch scanning, best-frame selection, and duplicate detection — run entirely on the device through Google Play services .

This architecture has two practical benefits. First, it delivers extremely fast performance because the computer vision and machine learning models are not waiting for a round trip to a cloud server . Second, and more importantly for documents containing sensitive information, your scanned images are never uploaded to Google’s servers during the capture and analysis phase. Complete offline functionality makes the scanner reliable in basements, airplanes, or anywhere with spotty connectivity . Google’s own documentation for its document scanner API confirms that this on-device model is designed to preserve user privacy .

The 8GB RAM hardware requirement

The convenience of on-device AI comes with a strict hardware filter. Google has confirmed that the new intelligent scanning features require a device with a minimum of 8GB of RAM to ensure smooth performance . This requirement is necessary because the machine learning models for real-time computer vision and frame analysis demand substantial memory to operate without lag .

The 8GB threshold effectively restricts the full experience to mid-range and flagship Android phones released in the last several years. If your device falls under that limit, you will continue to see the classic manual scanner interface. This hardware gate was observed throughout the beta testing period and is now a firm requirement for the stable public release .

A visual refresh and a year of steady improvements

The AI intelligence arrives packaged inside a completely redesigned visual interface. Google has applied its Material 3 Expressive design language, which introduces a rounded viewfinder, a redesigned tool carousel, larger page previews for easier editing, and a new thumbnail view that lets you reorder pages by dragging and dropping .

The current wide rollout on May 29-30, 2026, is not a sudden launch — it represents the final public release of a feature set that has been steadily built and tested in public for over a year .

• September 2025: Google redesigned the post-capture editing screen with Material 3 Expressive principles, adding larger page previews, improved crop and rotate tools , and new filters for cleaning up shadows .
• September 2025: A parallel test began through Google Workspace Labs for an "autopilot" scanner that automatically captured documents without requiring a shutter tap. This early experiment provided the foundation for the final Smart Batch Scanning behavior .
• Late 2025 – early 2026: The beta expanded to include deeper on-device AI integration. Users could toggle between the classic manual interface and the new automatic scanner, allowing Google to gather performance data and validate the 8GB RAM minimum before finalizing the stable release .

This incremental approach means the scanner you see today has been hardened over multiple release cycles. What began as a basic editing UI refresh last September has evolved into a fully automated, computer-vision-driven scanning flow that leaves the old tap-for-each-page workflow behind.