Google's Gmail Upgrade—You Must Decide If This Goes Too Far

Forbes

07-05-2025

Decision time for 2 billion users. NurPhoto via Getty Images
Google is changing Gmail. Putting aside the procession of recent attacks, some of which seemed to come from Google itself, the biggest threat could come from within. This leaves 2 billion users of the world's most popular email platform with a decision to make — and that decision is getting more critical and more difficult.
We're talking AI, and the accelerating updates to Gmail — and other platforms — as new AI features are added into the mix. We know that AI is driving a new tidal wave of threats from scams, malware and phishing attacks. But what we don't yet know is how safe and secure the exposure of personal data to corporate, cloud-based AI processing will turn out to be, once it has all bedded down and been exposed to leaks and attacks. Forbes Delete Any Texts On Your Phone That Include These Messages By Zak Doffman
Google has confirmed its latest AI upgrade is now here. 'Responding to email is now faster and better with Contextual Smart Replies," the company says. These are 'powered by Gemini' and 'generate even more detailed and relevant replies, based on the context of the email thread, ensuring your reply is addressing the issues at hand.'
Absent end-to-end encryption, which doesn't work outside walled garden platforms or enterprise systems, email is not an inherently private and secure medium. And so you should be wary of what you send. But even so, AI reviewing an entire email thread — potentially multiple threads — to formulate a smart reply takes us to a new level.
Google explains that 'if you're short on time or need help finding the right words, Gemini can analyze the context of an email and offer more detailed responses to fully capture the intent of your message.' You will be given a choice of what to send, after Gemini has 'taken the full content of the email thread into consideration.'
Contextual replies Google
This is available for Workspace Business and Enterprise editions, and Google says that it can be enabled by Admins in Gemini feature settings and by end users enabling 'smart features and personalization."
Google also cautions that 'Gemini feature suggestions don't represent Google's views, and should not be attributed to Google," that users should not rely on Gemini features as medical, legal, financial or other professional advice," and that 'Gemini features may suggest inaccurate or inappropriate information.'
This is clearly clever and helpful, but it should prompt some thought as how far is too far. Google is a corporation, and the fear that its platforms — and others — will now use AI to scrutinize private, confidential, secure information warrants reflection. Forbes Has Your Phone Been Hacked—This New Update Lets You Know By Zak Doffman
Google provides opt-outs and privacy guidance for all these features. It's down to you and the companies you work for to decide what happens next.
As I've noted before, there's an interesting twist to this debate. Google's addition of quasi end-to-end encryption to Gmail stops AI working on those emails. This even includes its new AI driven relevancy search. That's because Google (rightly) can't see encrypted emails. That neatly frames the debate.
Privacy or not — it's decision time.