OpenAI launches Atlas browser and uses it to train their AI models

[smallwheels]

OpenAI has just launched it's own web browser named 'Atlas' with integrated Chat GPT. That was to be expected, after Google has integrated Gemini with Chrome, Microsoft has integrated Copilot with Edge and Perplexity has also launched it's own browser 'Comet' back in July. Also vendors like Brave and Mozilla integrated AI models into their browsers.

However: It seems that per default Atlas is set to use the webpages users browse to train OpenAI's AI models, as theRegister describes in their (heavily recommended) Article about the new browser:

OpenAI puts ChatGPT into Atlas browser in bid to rethink web

: Why experience the web for yourself when there's so much privacy to surrender?

www.theregister.com

This is very bad news for forum owners that try to get rid of AI bots: If AI companies now start a distributed apporach, using genuine users to innocently train AI models while browsing the web, forum owners have no choice but to lock out those browsers completely. Which is probably the only working way to avoid getting your forum content to be sucked into OpenAIs models.

I do not now how Atlas identifies itself in the logs - if it does in a solid way it should be easy to block it in .htaccess or to redirect it to a page that says "you suck - we don't accept the usage of Atlas browser in our forum as it misuses it's users to secretly trains OpenAIs AI models." or something alike. It can only be hoped that many forum and webpage owners follow an approach like that.

On a sidenote: If you look at the TOS on the webpage of OpenAI they state:

What you cannot do. You may not use our Services for any illegal, harmful, or abusive activity. For example, you may not:

• Use our Services in a way that infringes, misappropriates or violates anyone’s rights.
• Modify, copy, lease, sell or distribute any of our Services.
• Attempt to or assist anyone to reverse engineer, decompile or discover the source code or underlying components of our Services, including our models, algorithms, or systems (except to the extent this restriction is prohibited by applicable law).
• Automatically or programmatically extract data or Output (defined below).
• Represent that Output was human-generated when it was not.
• Interfere with or disrupt our Services, including circumvent any rate limits or restrictions or bypass any protective measures or safety mitigations we put on our Services.
• Use Output to develop models that compete with OpenAI.

So basically, it is forbidden to use OpenAI's services to do what it's browser does per default and was OpenAI and other AI companies have been doing for years.

Hopefuly someone sues them heavily for what they do.

 

[Alpha1]

Even though they denied it in the past, Google has always secretly used Chrome as spyware in the sense that they tracked user behaviour and used that data for their products, including Search. Adding in AI to train on what users are browsing is a next logical step. We cannot block Chrome users from forums, so our content will be used to train AI.
Atlas is not popular yet. Chrome is.