Google Denies 'Misleading' Reports of Gmail Using Your Emails To Train AI (theverge.com) 37
An anonymous reader shares a report: Google is pushing back on viral social media posts and articles like this one by Malwarebytes, claiming Google has changed its policy to use your Gmail messages and attachments to train AI models, and the only way to opt out is by disabling "smart features" like spell checking.
But Google spokesperson Jenny Thomson tells The Verge that "these reports are misleading -- we have not changed anyone's settings, Gmail Smart Features have existed for many years, and we do not use your Gmail content for training our Gemini AI model."
But Google spokesperson Jenny Thomson tells The Verge that "these reports are misleading -- we have not changed anyone's settings, Gmail Smart Features have existed for many years, and we do not use your Gmail content for training our Gemini AI model."
What they didn't say (Score:2, Interesting)
Re:What they didn't say (Score:5, Informative)
Notice they said absolutely nothing about using it to target keyword ads at you, build profiles about you to target you with ads
Of course they didn't say that. They've always been open about doing that for unpaid consumer accounts, it's how they can provide the service for free. If you don't want your the ads, or for your data to be used, you can get that, starting at at $7 per month [google.com].
Re:What they didn't say (Score:4, Insightful)
They've always been open about doing that for unpaid consumer accounts, it's how they can provide the service for free.
An everybody else who provides a similar service for free does the same thing, only with worse spam filters (say what you want, Gmail has the best spam filters in the business - they don't want the competition).
Re: (Score:3)
a similar service for free
Free is a fiction that exists exclusively in the minds of stupid people.
Re: What they didn't say (Score:2)
Well copyright and privacy are dead for all but the incorporated.
And I wouldnâ(TM)t bank on a paid email account not being used for AI scraping.
There are no standards any more. Not even much of a pretense.
Re: (Score:2)
And I wouldnâ(TM)t bank on a paid email account not being used for AI scraping.
In Google's case, they're under quite a lot of FTC scrutiny, operating under two consent decrees, and they have an employee population that isn't known for keeping their mouths shut. It's possible that Trump's FTC might not act if he were paid off, but a leak would definitely generate a lot of press.
Re: (Score:1)
I don't know why you think this is a revelation. They have explicitly said they do this in the past. Why would you expect them to mention it now? This story is about AI training, no one asked Google for an itemised list of what they do.
Re: (Score:2)
They are personalizing ads with e-mail content for decades. And they are probably now using AI for that, who would have thought ... this does still not imply that an AI is trained with that e-mail or the advertiser gets to see the keywords. An advertiser may still guess the kind of e-mails you get when you click the ad, because they know what keywords they targeted.
Re: (Score:2)
Re: (Score:2)
Exactly. A classic deceptive answer, which implies that of course they are using emails to do some sort of training.
Re: What they didn't say (Score:2)
Re: (Score:2)
A century ago we learned how to legally lie:
we do not use your Gmail content for training*
Re:What they didn't say (Score:4, Interesting)
I just use GMail as a trash box for sign-ups. I don't give our my real address to anyone except friends and family.
They are welcome to plumb the depths of spam hell all they want...
Re: (Score:2)
Very specific denial (Score:3, Interesting)
"and we do not use your Gmail content for training our Gemini AI model".
What about other, or upcoming, models?
Re: (Score:2)
And what jurisdiction? If they are using European's emails to train AI models, it's going to be the mother of all GDPR fines.
I think that alone is probably a good enough reason for them not to do it world-wide, because even Americans get emails from Europeans, and the possibility of some of that information getting into their AI models and appearing in the output is a huge legal risk.
What's that saying again? (Score:3)
"Never take any speculation as being confirmed until a statement of denial about it is issued."
Re: (Score:2)
"Never take any speculation as being confirmed until a statement of denial about it is issued."
In this case a false denial would put them in violation of two FTC consent decrees, and would almost certainly leak (Google employees are not known for keeping their mouths shut), so it would be a particularly stupid thing to do.
Doesn't matter (Score:2, Troll)
It doesn't matter what they say, just disable them.
I'm a long time Gmail user, and I disabled these smart features years ago because they were annoying. I just want to see my inbox, sorted chronologically. Luckily they still let you do that.
However I continue to see them dangle new features in front of my face with the text "Turn on smart features to use this." They really want me to turn them on for some reason, I wonder why.
Re: (Score:2)
Re: (Score:2)
But I did turn off all that stuff because, meh, who knows, corporations are not known for their honesty.
AND I also have a Proton account for stuff that actually matters.
Re: (Score:2)
Re: (Score:2)
I believe them (Score:4, Insightful)
we have not changed anyone's settings
Because those settings have always done this
we do not use your Gmail content for training our Gemini AI model.
Because we're already done with that, but we will continue to sell it to advertisers because is our core business and always has been.
article updated (Score:3, Informative)
Plural to singular (Score:1)
Internet: "Google is using your email to train AI models!"
Google: "We don't use them to train our Gemini model."
Um... can we have more details? We suspect a devil infestation.
Saying the quiet part out loud (Score:2)
Google Denies 'Misleading' Reports
Translation: "Don't believe your lying eyes".
Your data used for other AI models, ads, ... (Score:2)
Re: (Score:2)
Google is a Privacy Rapist (Score:2)
Face it: Google is a data nympho.
I can them (and Meta[stasize]) "Privacy Rapists".
As usual (Score:2)
It's probably as usual: A company wants to write "Don't sue us if we use your e-mail with an AI model if you click the 'rewrite with AI' button," the lawyer writes future proof "Don't sue us when we use your data with AI" and the user reads "We will use your e-mails to train our AI to reproduce your e-mails!"
The only thing I am still missing is people posting on their walls (or whatever that is in the Google ecosystem) that they opt-out of the ToS.
Professional liar claims to be telling the truth (Score:2)
They won't get much out of mine (Score:2)
Not Gemini, how about other models? (Score:2)
Our Gemini AI Models (Score:1)