OpenAI has unveiled its own browser, ChatGPT Atlas. What can it do?

OpenAI has unveiled its own AI browser, ChatGPT Atlas.

The company announced this during a live broadcast.

The Atlas browser is designed for more personalized interaction with the internet. It can book tickets, edit documents, and perform other actions on behalf of the user.

When viewing any page, an “Ask ChatGPT” button appears, which opens a sidebar for working with content. For example, a user can open a movie review and ask the AI to summarize it. 

Interface of the overview page based on search results in Atlas / Screenshot 

The first platform on which Atlas is already available is macOS. The browser will subsequently be released for Windows, iOS, and Android.

What the page with images that the user is searching for in Atlas looks like / Screenshot

The browser allows you to literally “communicate” with the web by integrating ChatGPT into the process of viewing websites.

During the demonstration, OpenAI showed how the browser independently performs actions in “Agent Mode” — with the user having the option to press the “take control” or red “stop” buttons.

Agent Mode — which allows AI to perform actions autonomously (such as booking tickets or searching for information) — is currently only available to ChatGPT Plus and Pro subscribers.

A browser mode in which it searches for and books goods on behalf of the user / OpenAI

You can also select incognito mode in your browser or use your ChatGPT account.

Incognito mode in Atlas / OpenAI

According to Axios, the launch of Atlas could intensify the battle over how people — and their artificial agents — interact with the internet.

What Atlas cannot do

OpenAI emphasizes that the content you view is not used to train models unless the user agrees to this in the settings.

The company has also limited the agent’s capabilities:

• It cannot run code or install extensions.

• It does not have access to other applications or user files.

• It pauses to verify that the user is in control.

• It pauses on financial or sensitive websites to avoid accidental actions.

Browser scanning a document / Screenshot

At the same time, OpenAI acknowledges that even with all the protections in place, the risks of malicious instructions remain, so it advises users to closely monitor the agent’s actions and, if necessary, use the mode without logging into their account.

Competition among AI browsers

As Bloomberg notes, OpenAI has joined a number of companies that are trying to rethink the browser with the help of AI.

In particular, Perplexity AI recently launched its own assistant, Comet, while Opera and The Browser Company are integrating more and more artificial intelligence features into their products.

Following the news of Atlas’ launch, Google’s shares fell by almost 4.8%.

Back in September, Google integrated its Gemini AI model into Chrome. It can explain the content of web pages, summarize information from multiple tabs, and even restore closed sites.

Read more: X launches a marketplace for inactive nicknames. How will it work?