Updated on Nov. 30 with new analysis into Google’s AI upgrades and the critical decisions 2 billion users now face to maintain their security and privacy.
There’s nothing to worry about, Google tells Gmail’s 2 billion users, everything is fine. Unfortunately, that’s not the case. There is a serious issue that you need to worry about. All the misleading headlines and quickfire denials have not made that go away.
In recent weeks we have seen two versions of the same pattern. A loosely framed Gmail story triggering wild headlines and a public correction. First was a huge security breach framed as a new Gmail password leak. Then came a new Google policy on AI training.
The breach was not new and it was not linked to Gmail. An amalgamation of prior data leaks will always contain plenty of Gmail data — it’s the world’s largest email platform. And there has been no policy change on AI training on Gmail inboxes.
Google pointed all this out, and the wild headlines morphed into “nothing to see here.” This pattern will continue to repeat. Nothing has fundamentally changed.
But this is all bad news for users — the stories and the corrections. It points to a lack of understanding of privacy policies and what’s actually being done with user data. It’s not a Google-specific issue. But Google is the gorilla in the cage. And that matters.
Here’s the crux. Gemini is not your close personal friend, it’s not your confidente. Neither is ChatGPT or Copilot. They’re someone else’s computer, operating out of grey data centers, consuming an increasing percentage of the world’s energy and water.
You are responsible for your own privacy and security. Not Google. Not Meta. Not OpenAI. Not even Apple. You can select the vendors and platforms you think have the best architectures and defenses and privacy polices. But the choice is yours.
Google is often painted as a bête noire when it comes to privacy. But its platforms dominate. Just look at Chrome. An unassailable install base despite almost continual privacy warnings — including from Apple and Microsoft. This perfectly illustrates the disconnect. Users can’t say they weren’t warned when tracking takes place.
And so it is with Gmail. You have one very specific choice to make, as confirmed by Google repeatedly and again recently. Using its cloud-based services is your decision. Your opt-in. And even if some users have been automatically opted in by default, a quick check and two taps/clicks and that can be easily corrected.
If you allow Gemini — aka Google’s vast array of power-hungry servers in global data centers — to pore over your inbox, to analyze your private emails, however sensitive they are — then that’s fine. As long as you have made a conscious decision to do so.
Make a choice. Do not sacrifice your privacy through inertia or ambivalence.
And that’s where Google (and Microsoft and Meta and others) can be criticized. Privacy policies are a mess. It should be impossible for users to open their data to cloud services without a clear understanding as to what that actually means.
Meanwhile, tech giants will continue to pour billions into new AI capabilities and the wall-to-wall marketing shaping this new space race. As they do so, 2 billion Gmail users and the billions on all the other platforms will continue to sleep walk onto thin ice.
A new report puts Google’s unstoppable AI push across Gmail and other Workspace platforms into context: “Could Microsoft walk away with the corporate AI market?”
Josh Bersin suggests “Microsoft is staking out the market for corporate productivity and AI infrastructure, leveraging its massive Microsoft 365 install base and deep relationships with IT.” And “for individual productivity, Microsoft has now “embedded Copilot agents into Word, Excel, and Powerpoint, with insights into Outlook.”
Meanwhile, Parmy Olson writing for Bloomberg suggests “ChatGPT still has a winning edge over Google’s AI.” But what OpenAI lacks is a legacy install base. That makes Google and Microsoft hard to beat — ultimately.
The lesson from the recent launch of AI browsers, which don’t yet make much sense and so won’t yet disrupt the status quo, is that AI is at its most powerful when applied to what we’re doing already. Yes, ChatGPT defined a new category, but will this really succeed standalone or will it need to be stitched into Google’s or Microsoft’s productivity apps or into Apple’s or Google’s or Microsoft’s OS ecosystems?
In reality, we’re at the beginning of the space race. But what’s also clear is that privacy and user data security is not the differentiator we thought when Apple launched Private Cloud Computing and tried to use the iPhone enclave to expand into AI. Just as with browsers, users are showing they don’t yet care or worry enough about privacy. And even data security is something of an afterthought. Who doesn’t want Nano Banana?
Make no mistake — you’re already out on that thin ice.