Gmail has changed—here’s what you need to do.
Republished on February 4 with a new report into the illicit use of AI tools on sensitive data, and a raft of bans driven by China’s latest AI threat.
Be warned — if you don’t get on top of the AI being stitched through your apps and platforms, you will have no control over what data is being shared and who it’s being shared with. That’s behind the raft of new Gmail warnings, with millions of users now caught in a new world that few understand and that seemingly evades controls. Here’s what you need to do now.
With Google’s recent upgrade to add default AI settings to key Workspace apps — first and foremost Gmail, those wanting to safeguard their data are looking for easy-button controls. But right now, that’s a major issue. As 9to5Google explains, “as Gemini is now readily available to Google Workspace subscribers, admins have found that the ability to disable the new AI features is hidden behind a conversation with Google support.” Not good.
Data leakage through AI engines is now headline news courtesy of DeepSeek, China’s answer to Gemini and ChatGPT. And while that example is clearly acute, with user data and tracking heading to China by default and weak security controls, Gemini has also come in for criticism. “Opting out of Gmail’s Gemini AI summaries is a mess,” 404Media has warned, advising users “here’s how to do it, we think.” Again, not good.
As one Redditor complained, “I was annoyed by the new Gemini for Workspace features. In the admin page –> generative AI –> Gemini for Workspace, there were no settings to toggle on / off. I chatted with customer support, and at my request they ADDED the settings, after which I was able to toggle off all Gemini for Workspace settings (gmail, etc.).”
9to5Google explains that “it’s quite frustrating that, for a number of Workspace admins, the ability to turn off Gemini AI features appears to just not exist by default. Google, on a support page, says that the ability to ‘manage access to Gemini features in Workspace services’ is only available to ‘Enterprise Standard and Enterprise Plus’ accounts. That appears to be true, as a number of admins have found that the Workspace Admin Dashboard shows no option to turn off Gemini features. Some Enterprise users claim to not have these settings at all too… Customer support representatives apparently have the ability to enable a new menu of settings which can be used to turn off Gemini on the Google Workspace Admin Console.”
The AI data leakage was behind the DeepSeek scare, when the security industry pointed out that all the data shared with China’s disruptive AI engine was (unsurprisingly) heading straight for China. But as I have said before, while the Chinese angle has intensified the debate, it didn’t create it. The risk is still there, even with U.S. and other western engines.
Harmonic Security warns in a new report that generative AI tools “come [with] significant risks related to data security. Despite their potential, many organizations hesitate to fully adopt Al tools due to concerns about sensitive data being inadvertently shared and possibly used to train these systems. Organizations risk losing their competitive edge if they expose sensitive data. Yet at the same time, they also risk losing out if they don’t adopt GenAl and fall behind.”
The scale of this threat has now been reported by BBC News, which has looked into “why employees smuggle AI into work,” disclosing the sheer scale of a problem that now sees “many people using their own AI tools at work, without the permission of their IT division.” The BBC cites research from Software AG, which suggested that an alarming 46% of employees “would refuse to give them up, even if their organization banned them completely.” And when that AI accesses your email, the data exposure is huge.
We are also now starting to see the delayed response to DeepSeek’s sudden appearance as the next big thing in AI, with countries, companies and even states banning the platform from their sites and devices. This is a taste of things to come, and gets to the central issue of AI being users within organizations with little oversight.
For its part, Google assures users that with Gemini “your data is your data, the content that you put into Google Workspace services (emails, documents, etc.) is yours. We never sell your data, and you can delete your content or export it.” The company also says that “we do not use your Workspace data to train or improve the underlying generative AI and large language models that power [Gemini], Search, and other systems outside of Workspace without permission.” But while the company also says “they aren’t just words, to ensure we continually meet these high standards, independent auditors validate our practices against international standards and best practices,” is hasn’t yet assuaged concerns.
And that won’t be helped by Google’s confirmation that Gemini is now being misused by nation state threat actors to enhance and hone their attacks. Assuringly when it comes to Gmail, Google also says those “threat actors attempted unsuccessfully to use Gemini to enable abuse of Google products, including researching techniques for Gmail phishing, stealing data, coding a Chrome infostealer, and bypassing Google’s account verification methods.” But we’re still at the very early stages of this.
The concern is that when AI — by its very nature — is allowed inside the fence, it becomes hard to control. We don’t yet have the common transparency we need to be able to apply common sense security to the use of AI and ensure it sticks.
This is not isolated to enterprise users. The entire AI ecosystem now being added to the critical platforms and services we use needs a clear set of controls and policies. Instead it is evolving in real-time with little consistency. As 404Media says, while some users will “love” the new AI features, “others are going to want to run away from it as quickly as possible. Many people—including us—are already furious that they were automatically opted into it. Turns out, disabling it isn’t straightforward.”
In response to the furor on this, Google is keen to emphasize its privacy hub and the commitments it has made re user privacy. It’s clearly easier to opt out user by user (turning off “Smart features in Google Workspace”) than adopt a blanket approach across an enterprise, which does point towards the admin console per the questions that have been raised by users online.
Whether you’re a home or enterprise user, take some time to decide what’s right for you (or those you’re responsible for) and adjust your settings accordingly. Don’t be bounced into taking risks you’re not comfortable with by a mix of hype and FOMO. Harmonic Security says more than 80% of companies now worry about sensitive data leaking via generative AI.