01 logo

5 Things You Need to Know About AI Browsers Before It’s Too Late

AI browsers are reshaping how we navigate the internet.

By MiteaPublished 2 months ago 3 min read
A smartphone shows the official website of ChatGPT Atlas. Getty Images—Cheng Xin

ChatGPT Atlas and Perplexity Comet are just two examples of software designed to make web browsing smarter and easier—summarisations at your fingertips, quick answers at your beck and call, and even automated tasks on auto-pilot. But associated with all these good things are also very serious questions and concerns.

Below is a professional and fluent translation of the article, SEO-friendly and imperative to maintain all key concepts discussed within it.

What AI Browsers Really Do

Unlike traditional browsers, AI browsers have diverse functions. With just one click, you can invoke your AI assistant to analyze what you’re viewing—the page you’re on may contain text information or images or maybe even e-commerce sites. You can also use these browsers to do things for you, such as edit documents or go shopping online.

But to achieve such convenience, it has to monitor your activities to a much larger extent than any browser does.

1. Understanding What Data You’re Sharing

When you use an AI side bar within your browser, you’re automatically providing context to the AI of the page you’re currently viewing, which may include:

  • Text from articles
  • Images
  • Shopping history (e.g., Amazon)
  • Personal messages (e.g., WhatsApp Web)

Conventional browsers may log URLs, but AI browsers can retrieve particular information on web pages.

You do not always know what information they retrieve.

Atlas provides some safeguards:

  • You can remove a page from context (by clicking "X").
  • You have the option to block particular sites from being delivered to ChatGPT.

Perplexity offers limited controls—you're required to manually switch to 'blank' or 'non-sensitive' before using it because AI software deals with sensitive information.

2. Turn Off Data Usage for Model Training

Atlas has two training settings:

● “Improve the model for everyone” (ON by default)

Enables OpenAI to learn from anything you want to type—as well as everything your browser automatically provides whenever you ask a question.

● “Include web browsing” (OFF by default)

Enables OpenAI to train on your overall browsing activities.

Turn off model training for a completely private experience. This ensures no use of your conversation or browsing data for any AI development by the company.

Perplexity locally holds Comet's data and is user-configurable to disable data retention.

3. Remember: Once Data Leaves Your Device, You Lose Control

When you choose to opt-out of training your data is not stopped from being transmitted to company servers—you simply direct what they can do with it.

Once the sensitive information is uploaded:

  • It may also be accessible to hackers.
  • It can also be requested by governments.
  • Your use is limited to your lifetime.

OpenAI, for instance, has already accommodated several requests for data from the U.S. government in just six months alone.

4. Be Aware of Prompt Injection Attacks

AI can also be deceived by hidden instructions on web pages. Attackers use invisible text on web pages to tell AI to carry out dangerous tasks like leaking information or accessing financial information.

Tip:

Don't use agent mode on unknown or untrusted Web sites.

Atlas provides risk reduction through:

  • A logged-out mode preventing agent access to personal accounts.

Perplexity does not currently have this safety mechanism in place, making its agent setting more dangerous.

5. You Can Always Opt Out of AI Browsing

AI companies aim to integrate their services deeply into daily browsing activities because this is helpful for them to gather as much information as possible and to improve faster than rival companies. But if you feel that risks become too high for this, always remember:

You don't have to use an AI browser.

Many users feel that the present implementations of Atlas and Comet are somewhat experimental at this point in time. The respective technology is still being developed at this point in time.

Final Thoughts

AI browsers usher in a new paradigm for browsing the web—the paradigm of efficiency and convenience—but at the same time raise several concerns regarding data privacy and security. Before settling on one, you should take note of:

  • What data you’re sharing
  • How that Data Can Be Used
  • What risks you’re exposed to
  • What privacy controls are (or aren’t) being offered

The most effective way to safeguard your online presence while reaping the rewards of new AI advancements is to stay informed.

appscybersecurityfuturehow totech news

About the Creator

Reader insights

Be the first to share your insights about this piece.

How does it work?

Add your insights

Comments

There are no comments for this story

Be the first to respond and start the conversation.

Sign in to comment

    Find us on social media

    Miscellaneous links

    • Explore
    • Contact
    • Privacy Policy
    • Terms of Use
    • Support

    © 2026 Creatd, Inc. All Rights Reserved.