Can I use ChatGPT Atlas and other similar AI browsers for work?
Can I use ChatGPT Atlas and other similar AI browsers for work?
Embedded Files
You may have seen news about new AI-powered web browsers, such as the recently announced "ChatGPT Atlas" from OpenAI, Comet from Perplexity etc.
Due to data privacy and security risks, the use of the OpenAI Atlas browser (and any similar AI-integrated browsers from un-vetted vendors) is prohibited for any work-related purposes.
This applies to all company devices and any personal devices used to access company data.
The "Why": We Don't Have a Data Agreement
The "Why": We Don't Have a Data Agreement
This policy is a part of our existing Public Generative AI guidelines.
As our company AI policy states, you must not use personal, confidential, or sensitive company information with public AI tools like ChatGPT. The reason is that we do not have a Data Processing Agreement (DPA) with OpenAI.
A DPA is a legally binding contract that governs how a vendor (like OpenAI) can process, store, and protect our data. Without one, we have:
No control over where our data is sent.
No guarantee that our information is kept confidential.
No assurance that our data isn't being used to train their AI models, potentially exposing our trade secrets or client information to other users.
The New Risk: Your Browser Is an AI Agent
The New Risk: Your Browser Is an AI Agent
These new AI browsers are fundamentally different from standard ones like Chrome or Firefox. They are designed to be "agents" that don't just view web pages, they read, understand, and interact with them on your behalf.
Features in the new OpenAI browser, like "agent mode" and "browser memories," are designed to see your open tabs, remember your browsing history, and even take actions for you, like editing documents or summarising data.
If you use an unapproved AI browser for work, you are essentially creating a direct data pipeline from our internal systems to a third-party vendor we have no security agreement with.
Example: Logging into your Google Workspace, Salesforce, or any internal company dashboard through the OpenAI browser means everything on your screen could be seen, processed, and "remembered" by OpenAI. This falls foul of our data handling policies and puts company and customer data at risk.
Our Policy on AI Browsers
Our Policy on AI Browsers
These are covered by our Public Generative AI Policy and the following is prohibited:
Do Not Install: You must not install or use the OpenAI "ChatGPT Atlas" browser (or any similar non-approved AI browser) on any company-issued device.
No Work-Related Use: You must not use these browsers for any work-related tasks, even on a personal device. This includes:
Accessing your company email (Google Workspace)
Logging into our company services, intranet, or HR portals
Accessing any customer-related data
Accessing any internal code repositories
This policy also applies to any future competing products that offer similar "agentic" browsing features.
What You Need to Do
What You Need to Do
Only Use Approved Browsers: Continue to use our company-approved and managed browsers for all work.
Uninstall if Present: If you have already installed this browser on a company device for any reason, please uninstall it immediately.
Ask Before You Use: The world of AI is moving fast. If you see a new tool that seems useful, do not integrate it into your workflow. Please contact the the InfoSec or Desktop Support team before using or installing it.
We understand the excitement around new technology, but our primary responsibility is to protect our company, our employees, and our clients. Thank you for your cooperation in keeping our data secure.
Page updated
Report abuse