Google Workspace lets Gemini access your company data by default - how to shut it down
Follow ZDNET: Add us as a preferred source on Google.
ZDNET's key takeaways
- Internal AI access can create compliance and privacy risks.
- Workspace admins can turn off intelligence sources.
- Google gives Gemini access to Workspace data by default.
I have long accepted one fundamental truth: Google knows everything that I do. I've been using Gmail as my primary life dashboard for over a decade now. Apart from the occasional ad in my Gmail interface, Google really hasn't abused my trust. But still, Gmail sees all. Gmail knows all.
Every email message I have sent or received since before Hamilton opened on Broadway has been processed through Google's voracious data maw.
So, when it comes to the Gemini chatbot, I haven't been too upset that Google's AI seems to know what's in my email or my Google Docs. In fact, that capability helped me out at one point. I actually want Google's AI to do more, but the tech has a limited number of messages it can process. The AI crashed and burned on a more complex research task.
Also: Google is training AI on even more of your data now, unless you opt out - here's how
But not everyone wants Google to access everything in Gmail, Docs, Calendar, and Chat, particularly professionals who use the commercial Google Workspace service.
And yet, did you know that Google defaults to allowing Gemini access to all Workspace services? In this article, I'll show you how that process works. Then, I'll show you how to turn it off for your company.
What is this dark magic?
The good news is that while Google does automatically expose your company's Workspace data to Gemini, the tech giant doesn't use it for training or even use it outside your company's domain. Nor does Google share prompts or generated responses with other users or other organizations. It also doesn't create an AI-specific database from your documents and email messages.
Also: I tried Claude Cowork on my Gmail inbox after Gemini choked - and it saved me hours of work
This fact converts concerns around Workspace AI use from a potential five-alarm fire with deep regulatory and privacy implications to more of a corporate policy issue.
Gemini is available in Chat, Drive, Docs, Calendar, Meet, and Gmail. This capability means you can interact with the AI from whatever work surface you are currently using. To increase the relevance of AI responses while in those workspaces, Gemini accesses Workspace data as part of its response analysis.
Workspace considers each of these sources to be a Workspace Intelligence Source. On the surface, there's an interesting question you might raise. If Google considers each of these buckets a Workspace Intelligence Source but doesn't use them for training, how can it use them in AI queries?
Also: How to keep your conversations with ChatGPT, Gemini, Copilot, or Claude as private as possible
The answer is as old as Google itself: search. Google indexes all your Workspace sources (as it has since the beginning). When Gemini receives a prompt, the AI uses real-time Retrieval-Augmented Generation (RAG). Basically, Google searches the Workspace data as it would during search, retrieves relevant information it has permission to touch, and then formulates a query answer.
Internal data governance
However, just because your corporate data isn't shared with Gemini or used in its training, that doesn't mean you want it all available for AI use, even internally.
The most obvious reason you might want to restrict this capability is for compliance and regulatory reasons. You might have client contracts or regulatory restrictions that explicitly prohibit AI scanning and data retrieval from company information. You might also be subject to regulations that prevent data sharing, even internally.
Another reason is to enforce necessary departmental isolation. It's always possible that an AI could produce an answer that accidentally includes confidential records (think HR complaints, private deals, and more) and provide those answers to employees working in other departments.
Also: Why enterprise AI agents could become the ultimate insider threat
The same principle applies to the rising level of insider threats, both perpetrated by humans and by AI. The last thing you want is curious (or malicious) employees querying an AI to circumvent departmental safeguards to access information they shouldn't be privy to.
Even if you don't have curious employees, if Workspace data is automatically surfaced to the AI, innocent queries could give employees access to sensitive company info that happens to be living in a Google Doc or a Chat log.
How to shut it down
Log in to an account that has admin access to your Google Workspace. Then point your browser to admin.google.com. You'll see your main dashboard.
Click Generative AI (at 1) and then Gemini in Workspace (at 2). You'll see the Gemini in Workspace page. Now, click the Workspace Intelligence Sources block.
You'll get the individual settings page for the entire organization. Here, you can click an option to disable the source for the whole business.
I ran into an issue when I tried to turn off a feature for an individual user. You can enter a user into the users block (shown at 1). But when I tried turning off one of the sources, like Drive and Docs (at 2), there was no interface to disable the feature.
When I asked Gemini why I couldn't take that step, I was told it's because that selector is marked as view-only. To turn off features for individuals, those users would need to be moved to their own organizational units or added to a new group.
So now you know.
Will you let Gemini access your Workspace Intelligence Sources or shut it down immediately? Let us know in the comments below.
You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.