🔥 Trade with Pros on Discord → 21 Days Free (No Card)JOIN FREE

AI Giants launch browser war against Google’s Chrome

In this post:

  • OpenAI, Perplexity, and Microsoft have introduced AI-powered browsers to challenge Google’s 63% market dominance
  • Browsers provide AI firms with direct user access, valuable training data, ad revenue potential, and a platform for AI agents to handle tasks such as bookings and purchases.
  • Prompt injection attacks remain an unsolved threat, with Gartner recommending companies block AI browsers.

Top artificial intelligence groups are stepping up efforts to challenge Google’s dominance of the browser market, as they bet on the cutting-edge technology changing the way people access the internet.

OpenAI and Perplexity have both launched their own web browsers in recent months, while Microsoft has also rolled out its Copilot AI tool to its Edge browser, which enables users to ask questions to chatbots alongside the content they are viewing.

AI browsers will shape how “we’re all going to interact with the internet in the future”, said Mark Surman, president of the Mozilla Foundation, adding that his group planned to let users pick which AI model they wanted to integrate into its Firefox browser.

Browser control offers a path to users and revenue          

AI browser makers face tough competition from Google, which controls more than 63 percent of the global market share, according to Cloudflare, and has quickly incorporated its Gemini AI models into Chrome.

For both OpenAI and Perplexity, creating their own versions will give them a more direct relationship with their users, many of whom access chatbots such as ChatGPT on browsers controlled by Google and Microsoft.

Perplexity’s Jesse Dwyer, who works on its new Comet browser, said the start-up viewed browsers like the “operating system of your mind”.

Some have criticised the new AI-powered browser experience, with users complaining about glitchy and unreliable features. Others have warned about privacy concerns over personal data.

See also  New Atlantis in Starfield: A Hub of Exciting Side Missions

Google said it uses the conversations from Gemini to train its models, but does not harvest web page content from users’ browsers and removes personal information from browsing sessions.

OpenAI said its Atlas browser will follow existing training settings in ChatGPT. If a user chooses to opt in, that data is passed through privacy and safety filters before being used by the start-up.

Microsoft said its AI features are also “opt-in”, while Perplexity said it uses user data to train internal proprietary models that work on processes such as how queries are formulated, as well as to tweak frontier models such as ChatGPT, Claude, and Gemini to be more accurate.

Google’s lead will be hard to supplant, however. In May, it announced plans to launch “AI mode” in Google search and Chrome browser to provide a conversational, question-and-answer experience akin to OpenAI’s ChatGPT.

Its latest LLM Gemini 3, launched in November, is also considered to have leapfrogged OpenAI’s GPT-5 and achieved gains from the model training process that have eluded OpenAI in recent months.

“An AI-enabled browser alone isn’t a differentiator,” said Stephanie Liu, senior analyst at Forrester. “OpenAI will have to find a meaningful value proposition to pull in more users — which, again, is a tall order when they’re facing a very powerful, highly used incumbent web browser.”

See also  OpenAI releases Codex, its flagship AI coding agent, in ChatGPT

Adam Fry, OpenAI’s product lead for ChatGPT Atlas, said the company was working to bring more features, such as multiple profiles and the ability to group tabs, to its browser soon. “This is the start of a long investment we’re making in Atlas,” he said.

A big focus for Google has been developing agents and making capabilities such as translation and auto-filling information into forms much easier. It also launched a new experimental tool called Disco in December, which allows users to turn open tabs into custom, interactive apps.

Security risks emerge as models gain browser access

The integration of AI models into browsers also introduces new cybersecurity risks, such as prompt injection, where attackers can manipulate how LLMs behave by inserting malicious prompts on websites.

Prompt injection attacks are an unsolved security problem and stem from the fact that AI models cannot differentiate between legitimate user requests and malicious ones.

Consultancy Gartner recently suggested that companies block AI browsers over these cybersecurity concerns. These risks are exacerbated as users trust AI models with sensitive information such as credit card details. 

Despite the risks, most experts agree that browsers are ripe for innovation after largely remaining the same over the past two decades.

 

Sharpen your strategy with mentorship + daily ideas - 30 days free access to our trading program

Share link:

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Most read

Loading Most Read articles...

Stay on top of crypto news, get daily updates in your inbox

Editor's choice

Loading Editor's Choice articles...

- The Crypto newsletter that keeps you ahead -

Markets move fast.

We move faster.

Subscribe to Cryptopolitan Daily and get timely, sharp, and relevant crypto insights straight to your inbox.

Join now and
never miss a move.

Get in. Get the facts.
Get ahead.

Subscribe to CryptoPolitan