Above a personal email from a friend (name changed) in early April was a small summation of her short message to me from something called “AI Overview”:
- Barbara is getting a Jitterbug phone like Jan’s.
- Beth asked Barbara about her writing group…”
Below it in smaller type was the AI author’s message:” ‘By Gemini; there may be mistakes. Learn more”
There was no mistake. And I needed no translations of Beth’s email. Google’s Gemini program was invading my privacy. And because millions of others and I have written adversely about President Trump in our Gmails, we are now being watched. Could it be that the company is doing domestic spying for the government? God only knows what might happen from AI’s author summary of a lengthy email I had received from a fellow activist with this “overview”:
- George questioned weird AI summary and reported receiving strange email supposedly from you.
- George enjoyed dinner with mom; back hurting after gardening; concerned about getting sick again.
- George expressed strong opinions on recent political/religious news events.
Now, Google’s AI Gemini invasive program isn’t new, but the “AI Overview” feature certainly is. Its original purpose was apparently to summarize the main points in lengthy emails in business correspondence for busy users needing speed to reach the bottom line. Gemini also offers composition assistance (“Press / for help write”) for a variety of situations, from thank-you notes to job applications.
So it is impossible to believe that Google, its Gemini developers, and AI authors didn’t realize the program was violating the 4th Amendment’s privacy rights in our “paper.” Why else would President George W. Bush demand retroactive immunity for telecommunications firms hired to spy domestically in the FISA Amendments Act of 2008?
Instead, they seemed to count on public indifference and/or ignorance of the U.S. Constitution to escape what they thought would trigger only a few complaints about this significant violation. But thousands went quickly online with bitter complaints. I wasn’t alone. The powerful ACLU (American Civil Liberties Union) had already warned the public that: “government can create huge databases that contain information about U.S. persons obtained without warrants and then search these databases at a later point.”
The whole concept of spying on computer users’ messages via artificial intelligence started over two years ago with Meta (formerly Facebook). Its blue AI circle in its WhatsApp feature—used globally by more than 2 billion people —assured them about secrecy with end-to-end encryption, as well as “access, store, and analyze messages despite those in-app notices about privacy,” according to a massive class-action lawsuit filed January 23 of this year. It alleged that Meta “secretly allowed its employees, certain contractors and/or other third parties to intercept, read and store WhatsApp messages.”
Then came Microsoft’s Recall program, an upgrade to a Windows 11 feature via screenshots of every user’s past online activity. As its usefulness was explained:
“Recall was created to help users find anything they’ve ever seen on their PC. It improves upon most computers’ search functions, letting you scroll through a timeline of your past activity, from app usage to website visits. Because it takes screenshots of everything you do, you’ll always be able to find what you need.”
It wasn’t long before a costly ($100 billion in market value) error of fact was detected by many astronomers. Google hurriedly renamed the program “Gemini” for the star constellation of two mythical brothers. They were “closely associated with the free-flowing exchange of ideas, communication, and interaction,” according to the Farmer’s Almanac.
This Overview feature was quietly inserted into the Gemini program soon after, under the name “Workspace Smart Features.” Perhaps to avoid drawing attention to it, Google failed to ask for user consent directly.
A firestorm exploded over Overview’s Gmail summaries. Thousands wrote comments like these:
“I just got a gemini ai summary that read a confidential email about a relative’s medical diagnosis and poor treatment at an emergency room. This really upset me and I sent a blistering message to google complaining about the obvious Hepa violation.”
“When I asked about my first crush, Gemini was able to determine that it occurred in elementary school, as well as tell me the name of my first love, how we met, and when….How does Gemini know that? The simple answer is that the information is somewhere within the 16 years’ worth of my email history that it can access.”
“I was so ticked off to find an AI summary of emails between my brother and me! This is so freaking intrusive and obnoxious…but now I’m so aggravated with Google and concerned about their privacy policies that I plan to upgrade my free Proton mail account and migrate to that as my primary email address.”
Most of us just wanted to get rid of the Overview message summaries and were helped by experienced users’ instructions that boiled down to this:
- On your computer, go to Gmail.
- At the top right, click SETTINGS. Go to SEE ALL SETTINGS.
- Under the GENERAL tab, scroll to GOOGLE WORKSPACE SMART FEATURES.
- Click on MANAGE WORKSPACE SMART FEATURE SETTINGS.
- Turn a “smart feature setting” on or off, such as “Smart features in GOOGLE WORKSPACE”. Or Smart features in “OTHER GOOGLE PRODUCTS”.
- Click SAVE at the bottom right.
However, some of us have aging laptops that stopped at instruction No. 3, meaning the “overviews” will continue unless we switch from Google to other free email programs. Dozens exist.
One Illinois user was furious enough to file a class-action lawsuit (Thele v. Google LLC) last November in California’s Northern District Court against Google. It may well start a tsunami of privacy litigation. Demanding a jury trial, he’s alleging that:
“…on or about October 10, 2025, Google secretly turned on Gemini for all its users’ Gmail, Chat, and Meet accounts, enabling AI to track its users’ private communications contained in those platforms without the users’ knowledge or consent. As of the date of this filing, Google continues to track these private communications with Gemini by default, requiring users to affirmatively find this data privacy setting and shut it off, despite never “agreeing” to such AI tracking in the first place…Despite users never giving Google informed consent to track and store their private communications, Google does just that, using Gemini AI to access and exploit the entire recorded history of its users’ private communications, including every email and attachment sent and received in their Gmail accounts.”
Moreover, when Google activated Overview in November 2025, the tech media and The New York Times immediately pounced to warn users about its privacy violations (Inc.: “Cybersecurity Experts Warn That Most Gmail Users Don’t Realize This AI Setting Is Already Turned On” NYT: “A.I. Has Arrived in Gmail. Here’s What to Know”). Unfortunately, most Gmail users don’t have nor seek access to these publications—and Google didn’t explain the Overview choices.
Nor are they likely to look up the Fourth Amendment to see how they are protected by the courts from this invasion:
“The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.”
Google’s directors and shareholders need to resolve this black eye before millions of email users switch to other free and secure email services. There are several top-rated choices: Proton Mail, Tuta, Outlook.com, iCloud Mail, Yahoo Mail, Zoho Mail, Mailfence, GMX Mail, AOL Mail, and Mail.com.
Leading any decision is whether to stop using Overview summaries and any contractual domestic surveillance arrangements with U.S. governmental agencies such as the National Security Agency, the Federal Bureau of Investigation, and the Department of Homeland Security. Too, those firms are being paid with taxpayer dollars, something that can prompt a Congressional oversight hearing, with its destructive publicity for the companies involved.
Growing privacy litigation strongly suggests that federal judges and the U.S. Supreme Court will rule for the Fourth Amendment rather than for the FISA Amendments Act of 2008, which provides “retroactive immunity” to telecommunications companies hired to spy on the American public. Who else has the means and equipment to spy on us? And if those firms lose, it could cost them billions, especially from a class-action suit.
If Google and others continue to be Peeping Toms, taking the government’s “30 pieces of silver,” email users need to pay attention to email colleagues like “Shane,” who pointed out technology’s grievous untrustworthiness:
“As long as companies continue to introduce ‘features’ that are enabled by default and without a user’s consent, there will always be a privacy problem in tech.
As long as there are policy updates hiding invasive technologies that are enabled by your ‘consent’ to keep using what you already have, and without providing an opt-out, there will always be an ethics problem in tech.
As long as you have to dig through settings to find footnotes about aspects or features you must disable after-the-fact, or cannot disable at all and the damage has already been done, there will always be a trust problem in tech.
We call it zero trust for a reason.”