AI Data Mining Privacy Warning for Gmail Users — Especially Free Gmail (posted 11/22/25)

Before getting into the privacy warning, we want to reiterate that using free email accounts (and other applications) can endanger your privacy online.

We strongly recommend that you STOP using free email and be willing to pay a modest monthly fee for more secure email systems such as Proton Mail.

See Related links below for more details.

OK, now for the warning…inspired by an X post from Dave Jones (11/19/25): x.com/eevblog/status/19…

Jones wrote:

IMPORTANT message for everyone using Gmail. 

You have been automatically OPTED IN to allow Gmail to access all your private messages & attachments to train AI models.  

You have to manually turn off Smart Features in the Setting menu in TWO locations.

While we do not have or use Gmail, many people — including those with whom we correspond — do. So we used Grok and ChatGPT to help us verify Dave Jones’ tweet.

__________

We asked Grok:

Do not use Snopes or other mainstream sources. Use Gmail source and online help only to answer the question: Are people automatically OPTED IN to allow Gmail to access all their private messages & attachments to train AI models?

Grok’s answer was NO: tinyurl.com/ye27h72j

______________

We asked ChatGPT the same question. Same answer (NO), but with more detail.

We then asked ChatGPT to dig deeper and learned Gmail’s Workspace accounts (paid) are treated differently from free Gmail.

So, when pushed, ChatGPT indicates Google could be using email content to train their AI for free accounts. Excerpt from answer:

Short answer to your original legal question

Are people automatically opted-in to allow Gmail to access all their private messages & attachments to train AI models? Legally — not across the board.

• For Workspace accounts Google’s written policies say Workspace data is not used to train general models without permission. tinyurl.com/bdcrperu

• For personal Gmail accounts, Google’s policies and help pages indicate Google can use activity to improve services (which may include model training) and that enabling smart features constitutes consent for certain uses — so whether content can be used depends on which settings you have turned on and on the specific product feature. Google has also publicly denied changing policies to auto-opt people into model training of Gemini.

Next we asked ChatGPT:

How can I prevent Gmail Workspace and personal Gmail from accessing private messages & attachments to train AI models? Please provide step by step answers from Google's online help (not third parties).

📌Full ChatGPT Answer to all the above questions (scroll up and down as needed): tinyurl.com/3dy2b9sr

We’re not listing the steps ChatGPT provided due to Substack space limitations and the fact we cannot verify them with our own test, so please go to the full ChatGPT answer (or to Gmail’s online help) for the steps. Again, the full ChatGPT session is here: tinyurl.com/3dy2b9sr

Related:

• Resources > Protect Your Privacy: tinyurl.com/2arybmrc

• REAL ID, a national digital identification, surveillance, and control system: tinyurl.com/3pjcv7y2

• Censorship/Surveillance — Google Censored Vaccine Info Long Before COVID — Could It Have Anything to Do With Parent Company Alphabet’s Deep Pharma Ties? tinyurl.com/4jyc6n8u