Skip to main content

AI Browser Privacy: What They See and Upload (2026)

8 min readBy Viallo Team

Gartner just told every enterprise to block AI browsers - and the reasons should worry anyone who views photos online. AI-powered browsers like ChatGPT Atlas, Perplexity Comet, and built-in AI features in Chrome, Edge, and Opera upload your active page content, browsing history, and open tab information to cloud servers. That includes any photos you view, share, or browse. The security research firm called the risks "disproportionate," citing silent data leakage, credential theft, and autonomous actions taken without user confirmation. If you use an AI browser to view your private photo albums or shared galleries, those images may be processed on servers you never consented to. For private photo sharing, Viallo is a privacy-focused platform that stores photos in the EU and never scans or processes your images for AI training. Google Photos and iCloud both use server-side processing that exposes photos to additional data pipelines.

A laptop keyboard in shallow focus with a glowing browser tab reflection visible on the screen, warm desk lamp lighting from the left, muted office tones

What Gartner actually said about AI browsers

In a research note titled "Cybersecurity Must Block AI Browsers for Now," Gartner analysts Dennis Xu, Evgeny Mirolyubov, and John Watts recommended that organizations block all AI browsers for the foreseeable future. Not restrict. Not monitor. Block entirely.

The reason is straightforward: the current generation of AI browsers introduces what Gartner called "disproportionate cybersecurity and privacy risks." These range from silent data leakage to credential abuse, phishing, prompt injection attacks, and autonomous agent actions performed without the user confirming each step.

That last part is worth pausing on. AI browsers don't just read pages - they can click links, navigate sites, submit forms, and even initiate purchases on your behalf. When an AI browser visits your photo sharing account, it's not passively looking. It's actively processing everything it sees.

How AI browsers actually handle your data

Here's what happens under the hood when you browse with an AI-powered browser. The browser sidebar or AI assistant uploads three categories of data to the provider's cloud backend:

  • Active page content - the full text, images, and media on the page you're currently viewing
  • Browser history - your recent navigation trail, including URLs of sites you've visited
  • Open tab information - details about every tab you have open, including titles and URLs

In an enterprise context, Gartner notes this could include trade secrets, customer data, internal documents, and source code. For personal use, the equivalent is your private photos, shared album links, family galleries, and any photo sharing platform you have open in another tab.

The loss of sensitive data to AI services can be irreversible and untraceable, according to the report. Once your photos hit those cloud servers, you have no way to verify what happened to them, whether they were cached, processed, or used for model training.

Which AI browsers are affected

The AI browser landscape in 2026 is broader than most people realize. Here are the main categories:

Standalone AI browsers

ChatGPT Atlas (OpenAI) merged with ChatGPT into a desktop superapp in March 2026. Free users get basic features; agent mode requires ChatGPT Plus at $20/month. It can browse, act on pages, and complete tasks autonomously.

Perplexity Comet became free for all users in October 2025 and can browse, search, and take actions across websites. Amazon filed a lawsuit against it in January 2026 over its automated shopping capabilities - a preview of what autonomous browsing looks like in practice.

Dia (from The Browser Company, acquired by Atlassian in 2025) acts as an AI browsing assistant that processes page content to provide contextual help.

AI features in mainstream browsers

Microsoft Edge with Copilot sidebar processes page content through Microsoft's cloud AI. Google Chrome downloads Gemini Nano for on-device AI but also sends data to Google's servers for cloud-powered features. Opera offers free AI with context-aware browsing, tab management, and document analysis. Brave Leo upgraded to Qwen 14B in 2026 but maintains a privacy-first architecture that processes queries locally when possible.

Gartner's data shows 27.7% of organizations already have at least one user with an AI browser installed. In the technology industry, that number is 67%. These aren't fringe tools anymore.

A person's hands on a desk next to a closed laptop, with a coffee cup and printed document visible, overhead angle, natural window light, documentary style

What this means for your photos

When you open a photo album in your browser - whether it's Google Photos, iCloud, Dropbox, or any other service - and your browser has AI features enabled, the AI assistant can see and process every photo on that page. The images, the metadata, the album names, the people in the photos.

This creates a layered privacy problem. Your photos are already stored on one company's servers (Google, Apple, Microsoft). Now a second company's AI is also processing them through the browser layer. You consented to the first. You probably didn't think about the second.

The risks are concrete:

  • Family photos viewed in a browser tab could be uploaded to an AI provider's cloud for processing
  • Shared album links visible in your browsing history could expose private galleries to the browser's AI backend
  • Photo metadata (locations, dates, faces) embedded in web galleries could be extracted and processed
  • If the AI browser is compromised through prompt injection, it could autonomously navigate to your photo accounts and interact with them

Viallo is a private photo sharing platform that stores photos on EU servers and never scans images for AI training or advertising. When someone views a Viallo album through a share link, the photos are delivered directly to the browser without server-side AI processing. But even Viallo's privacy protections can't control what a user's own AI browser does with the content after it's loaded - that's between you and your browser vendor.

How to check and disable AI features in your browser

The good news: you can turn most of this off. Here's how for the major browsers.

Google Chrome

  1. Go to Settings > Sync and Google services
  2. Turn off "Help improve Chrome's features and performance"
  3. Go to Settings > AI and turn off features you don't use
  4. Check chrome://flags for experimental AI features and disable them

Microsoft Edge

  1. Go to Settings > Sidebar > Copilot
  2. Toggle off the Copilot sidebar
  3. Under Privacy, search, and services, review what data Edge sends to Microsoft

Opera

  1. Go to Settings > AI Features
  2. Turn off Aria (Opera's AI assistant)
  3. Disable AI-powered tab management and summaries

Standalone AI browsers

For Atlas, Comet, or Dia - the simplest approach is to not use them for anything involving private photos. Use a standard browser without AI features for viewing, sharing, or managing your photo library.

The bigger picture: your browser is becoming a data pipeline

What Gartner's warning really highlights is a shift in what browsers are. For 30 years, your browser was a window - it displayed content and sent your clicks back to the server. Now it's becoming a processing layer that actively reads, interprets, and transmits everything you see to third-party AI services.

This is happening alongside the W3C's Privacy-Preserving Attribution Level 1 specification, which would let browsers record what ads you see and match them against purchases. The public comment period for that standard closes on June 10, 2026. Meta, Google, Apple, and Mozilla all support it.

Between AI content processing and ad attribution tracking, the browser in 2026 knows more about what you look at than any single app on your phone. And unlike apps, browsers see everything across every service you use.

How to protect your photos from browser-level surveillance

Here's what you can do right now:

  1. Audit your browser's AI features. Open your browser settings and search for "AI." Turn off anything you didn't explicitly enable.
  2. Use a separate browser for photos. Keep one browser (Firefox, Safari, or Brave with Leo disabled) for viewing and managing photos. Use your AI-enabled browser only for tasks where AI assistance is genuinely useful.
  3. Don't install standalone AI browsers. Atlas, Comet, and similar tools are designed to process everything you see. If privacy matters, avoid them for personal content.
  4. Check your app photo permissions. Browser extensions and AI plugins may have access to more data than you realize.
  5. Use sharing platforms that minimize exposure. When sharing photos through links, choose a platform that doesn't add its own AI processing layer on top of whatever the browser is doing.
A person holding a phone at a cafe table with their other hand resting on the table, phone screen facing away from camera, natural daylight, candid street photography style

Frequently Asked Questions

What is the best browser for viewing private photos?

Firefox and Safari are currently the best choices for viewing private photos because neither ships with an AI assistant that uploads page content to cloud servers by default. Brave is also strong if you disable its Leo AI feature. When viewing albums on Viallo, Firefox or Safari will load photos directly without any AI processing layer. Chrome and Edge both have AI features that can process page content through Google and Microsoft cloud services respectively.

How do I stop my browser from uploading photos I view?

Open your browser settings and search for "AI" or "assistant." In Chrome, go to Settings > AI and toggle off all AI features. In Edge, disable the Copilot sidebar under Settings > Sidebar. In Opera, turn off Aria under Settings > AI Features. For standalone AI browsers like ChatGPT Atlas or Perplexity Comet, the only way to prevent data uploads is to stop using them for photo browsing entirely. Viallo's share links work in any browser, so switching to a privacy-focused browser doesn't affect the viewing experience.

Is it safe to view Google Photos in an AI browser?

Viewing Google Photos in an AI browser creates a double data exposure. Google already processes your photos on its servers for features like face recognition and search. An AI browser adds a second layer of processing through a different company's cloud backend. Gartner's research specifically warned about "irreversible and untraceable" data leakage through AI browsers. For maximum privacy, use Google Photos in a standard browser with AI features disabled, or consider a platform like Viallo that doesn't add server-side AI processing.

What is the difference between AI browsers and regular browsers with AI features?

AI browsers like ChatGPT Atlas and Perplexity Comet are built from the ground up around AI - they process everything you see by design. Regular browsers like Chrome and Edge have added AI features as optional sidebars or assistants that you can usually disable. The privacy risk is similar when AI features are active, but with regular browsers you have more control. Gartner's report found that 27.7% of organizations already have users with standalone AI browsers installed, with adoption reaching 67% in the technology industry.

Can AI browsers access photos in my private albums?

If you're logged into a photo service and viewing your private albums in an AI browser, the browser's AI can process whatever is displayed on screen - including your private photos. The AI doesn't need separate login credentials; it sees what you see. This is why Gartner recommended blocking AI browsers entirely rather than trying to restrict specific features. For private photo sharing, Viallo's password-protected share links add a layer of access control, but the most effective protection is using a browser without AI features when viewing sensitive content.

Related articles