Key conclusions:
- AI browsers collect huge amounts of data: They track your history of viewing, loading, search queries, open tabs and even confidential information.
- They can use this data for teaching AI and advertising: The collected data can help them train AI models, create detailed user profiles and advertise fuel.
- You can still protect your confidentiality: Use “excluding” modes, when available, turn off the memory functions, refused to teach data and avoid the exchange of confidential details that you do not want to open.
AI companies come to your browser.
Embarrassment launched its browser AIComet, in July of this year with a large number of fanfare. Openai too Soon launch your own browser AIThe field and anthropic has already introduced Browser -based agentClaude, in August 2025.
But after the post -end of AI companies can launch browsers with a large motive: not user performance, but user data.
According to the new study of researchers from UCL, UC Davis and Mediterranean University Revo Calabria, AI Web browser assistants represent significant risks of confidentialityField
Researchers found that these assistants collected and exchanged confidential user data, such as social insurance numbers and medical records, without adequate safety measures.
Read further to find out the risk of the confidentiality of AI browsers and how to protect your data when using it.
What these browsers AI collect
After the review Comet's confidentiality policyWe found that he collects and stores the following data:
- URL -address web pages that you visit
- Text, images and other content from these pages
- Permissions provided by web sites (for example, access to location, camera and microphone)
- The number of open tabs and windows
- Search for requests you enter
- File entries that you download from sites
- Cooks stored on the web sites that you visit
In addition, it can collect technical data, including the operating system of your device, the special support specification and your IP -address.
If the AI browser collects this amount of data, I will worry. What do they do with all this? And what will happen if the data is exposed in data violations?
The notification of confidentiality inaccessibility notes that he can share the collected data with potential buyers, new owners or other parties during large business changes.
CEO of Srinivas was indicated in authoritative media platforms, such as Techcrunch And MaysabricSupposing that he seeks to create a browser in order to track everything to sell hyperlinological advertising.
This is a kind of one of the other reasons why we wanted to create a browser is that we want to get data even outside the application in order to better understand you. We plan to use the entire context to create the best user profile, and perhaps you know, thanks to our Discover channel, we could show some ads there.
– Srinivas, TBPN Podcast
The General Director of the “Party” even made an application for $ 34.5 billion To purchase Google Chrome, a bold step from the company $ 18 billionAlmost half of the size of their goal.
Why? The answer may be simple: user data. Chrome teams 63.6% of the global browser marketImagine the volume of the user's behavioral data.
Another browser with AI, Microsoft, launched his AI -based mode AI In Edge Browser, which collects user data to improve the offer of their product.
Microsoft privacy policy It clearly says that he collects your data for work, updating and providing his products, offering support if necessary.
The company also uses your data to improve services, personalize experience, advertising and suggestions of special recommendations.
Microsoft – Microsoft, right?
Together, these examples show that AI browsers are not just viewing tools. They are also extremely effective data collection tools
Why are companies in artificial technologies tear out browsers from AI
Artificial intelligence companies are involved in launching browsers by AI, primarily to obtain control over user data. This is understandable how Artificial intelligence companies do not finish training dataField
According to Wall Street Journal, AI company I am looking for alternative sources of user dataFor example, the Openai field can use YouTube transcriptions for learning GPT-5.
AI browsers allow these companies to directly collect user behavioral data, such as searches, clicks and habits of content consumption. Then they can use this huge amount of data to teach their AI models.
In addition, AI companies can use behavioral data to create user profiles for target advertising. The creation and management of AI models requires a huge amount of money, so it is not surprising that Openai remains unprofitable to this dayField
Since advertising remains one of the most profitable digital businesses, artificial intelligence browsers offer a strategic method for these artificial intelligence companies to monetize their technologies on a scale.
Another reason why the AI companies strive to launch browsers in artificial intelligence is their strategic position in the market. The launch of company browsers allows them to diversify their products and compete with industry leaders such as Google.
However, with such deep data comes the real value: your confidentiality. AI browsers can easily become observation tools, quietly collecting detailed information about your online activity.
Researchers from UCL, UC DAVIS and Mediterranean University REGIO Kalabria said:
Our conclusions show that the assistants of the Genai browser collect and share personal and confidential information for profiling and personalization, emphasizing the need to protect guarantees, since they are increasingly mediated by a web examination.
Protecting your confidentiality when using AI tools
AI tools, whether it is a chat -bots or browsers, collect a huge amount of user data for teaching large language models that feed them.
As the art companies reach the walls in data training, the provision of teaching AI to user browsers is the next step to collect behavioral data.
For example, the “embarrassment” comet collects previous searches and retains memories of previous conversations in order to better answer your questions and provide “the best” experience.
But what if you do not want to train AI with your data or, even better, save your actions and preferences in viewing, how did the mother -Priro intended?
Here are a few steps that you can take to protect your confidentiality when using browsers AI:
- Try to use the version without taking into account where it is possible. This can limit the amount of personal information that can collect AI tools.
- Avoid using Facebook or Google. Think about using a fake email address to hide your real one. Proton Pass' Hide-My-Cheee-Email Acepses Great for this (albeit a premium function).
- Regularly delete your data if the tool allows you to do this in the “Settings” menu. In addition, turn off the memory function, if any (the comet has it).
- Abandon the training of AI When possible. Google allows you Manage data to which AI Gemini has access on your Android phoneField
- Do not share anything from AI that you would not want to expose in a violation of data.
Protecting your confidentiality in the AI era is a big problem. These tools increase your work process, research process and give you an increase in your domain, but often due to your data, confidentiality and, possibly, anonymity.
Popular browsers, such as Edge, already have a reputation as user tracking and creating profiles for target advertising. But AI browsers make further tracking users a few steps … much further.
Take proactive steps to protect your confidentiality when using browsers AI and other artificial intelligence tools – Your data should be yours alone, and you will never have too much confidentiality.
The editorial policy of Tech Report is focused on providing useful, accurate content, which offers real value for our readers. We work only with experienced writers who have specific knowledge on topics that they cover, including the latest developments in the field of technology, confidentiality on the Internet, cryptocurrencies, software and much more. Our editorial policy guarantees that each topic is studied and supervised by our internal editors. We support strict journalistic standards, and each article is 100% written by real authors.
