Google Update Reveals AI Will Read All Your Private Messages
forbes.com
forbes.com
"The Google Messages app creates suggestions with technology that’s built into your device. This way, your conversations stay in the Google Messages app. They aren’t sent to the Google Assistant or Google servers."
"There will be another, less contentious privacy issue with your Messages requests to Bard. These will be sent to the cloud for processing, used for training and maybe seen by humans—albeit anonymized. This data will be stored for 18-months, and will persist for a few days even if you disable the AI, albeit manual deletion is available."
This person shard a report on this issue for Ireland's Data Protection Commission: https://eupolicy.social/@thatprivacyguy/111839553765828301
And the person that shared "a report" only shared this same Forbes article.
2. Get the lie published in some sketchy media outlet that doesn't fact check
3. When challenged that your lie is a lie, cite the sketchy media outlet as your evidence
4. Mix in LLM hallucinations for more fun and profit!
Your new link is even worse. It just quotes from this article, but conveniently omits the part about the original source being a chatbot.
It's quite amazing how gullible some people are. I can kind of understand HN commenters reading the headline, believing it, and commenting without reading the article. But this is somebody whose self-identity is so tied to online privacy issues that their handle is "thatprivacyguy", and they're literally filing official complaints with no fact checking despite obvious red flags, and wasting precious DPC resources on AI-generated slop.
It tries to lead you into thinking that every RCS message is read by Bard on Google servers in clear-text, without e2e encryption. When it is only chats with the Bard chatbot that are readable by Google, for obvious reasons.
Then the article talks more about Apple than Google.
Also the author seems to be querying Bard for original information without double-checking.
This is the reason why Google is so hesitant, anything they do will be used against them.
Sorry for this bad comment, but this seems to be a Gish gallop situation so I want to get over it as quickly as possible.
As for what has actually been announced (LLM-based writing assistance within Google Messages), do people actually want chatbots to write their text messages for them? I certainly don't. But what a terrible world we'll be living in if "AI" anti-features like this take hold. Next to no real-world utility from LLM tech, but just one more layer of inhumanity injected into our worsening communication. This is a sure way to send messages you don't mean, but it's even worse because there's probably another LLM on the other end, summarizing all meaning and personality out of the messages anyway. So all text-based human communication will soon be reduced to a neverending string of misunderstandings thanks to the LLM-telephone game happening between you and your contacts.
As an iPhone user who has removed Google as much as possible out of my life, I have basically zero control over what happens with my messages when I send it to people on Android. Same with email, my contact information (particularly if they just allow any app to get their contacts), etc.
Put simply we are beyond the point that just because information is on your phone that it means all of it is "your" information and you alone can consent to it.
Particularly if it's sensitive texts, photos, whatever. But it shouldn't matter how sensitive it is.
I don’t trust Apple at all, naturally, but I possibly like them a bit more than I like Google.
Apple will likely try to violate our personal data and messages at some point if they see Google getting away with it. That’s how it works. One chip at a time. They might even be compelled to do it by law with a gag order.
However the contacts issue is a problem on iOS as well and I think it would be somewhat trivial for Apple to address that for any contacts that are associated with an iPhone user. Make it so the user can allow apps on someone else's phone to get their information (I am not implying this is easy, but some of the parts to facilitate this may already be there).
However it is also important to look at that Apple has been pushing the other way and working on implementations to make complying with those orders impossible. While currently iPhone backups are a major problem right now there have been multiple reports that they are working on E2E encryption of their entire iCloud system.
So at the moment I have far more trust in them considering their entire business model isn't built on gathering up as much information as possible.
That's really just an impression. First, you can use E2EE with other apps (e.g. Signal). But mostly, E2EE only secures the messages between the two devices, NOT on the devices. Apple, just like Google, basically owns your device. Whatever you can read on your screen, they can technically access, too.
Ever heard of Pegasus? Third parties can access smartphones with exploits like this, meaning that they just go around all the E2EE you can have.
There is a massive difference between exploits, gag orders, etc and normal day to day operations like what Google and Facebook are doing here.
I am not saying apple is saints, in my first line I very clearly said "companies" which includes Apple. But let's not pretend that Google and Facebook have not made Privacy invasion part of their business model and removed the ability for me to consent to my data being consumed by them regardless of how much effort I make to remove both from my life.
I don't. I also don't pretend that Apple is better, though it seems to me that you do :-).
> normal day to day operations like what Google and Facebook are doing here.
I don't think that Google or Facebook extracting everything that appears on your screen (including e.g. Signal messages) to their servers would count as "normal day to day operations", though. I would guess that it would even be illegal. However they can probably do stuff with the data that you send to their servers unencrypted (iCloud too), which is where E2EE is interesting. But then again, using Signal with a recipient on Android or iMessage with iOS won't make a big difference, I would say.
I trust then more, our house is full of Apple gear - but I still can’t trust them.
Re: Google. They weren’t always monsters, there was a time when they were positively viewed upon on the web pre-2006.
My key problem here is consent. I chose to use Apple, I theoretically can choose not to use Google but that doesn't mean anything since they will get my data anyways, at least some of it.
And agreed on Google, I used gmail, reader, and basically any experiment they came out with. For me it was somewhere around 2008 from what I can tell that I started moving away from them.
Both 3rd and 4th amendment apply.
"No Soldier shall, in time of peace be quartered in any house, without the consent of the Owner, nor in time of war, but in a manner o be prescribed by law." The soldier being Google and the house being my phone.
[0] https://en.wikipedia.org/wiki/Wikipedia:Reliable_sources/Per...
That saying predicts the future better than a time machine would.
9to5's leak shows an option of chatting to bard in google messages and this article takes that to make the point that all your messages will be shared with Google.
reference article: https://9to5google.com/2024/01/18/google-messages-bard/ This is more like those indie tools that send message to chatgpt from your texts app.
And that's gonna be not-encrypted anyway.
Doesn't desktop Outlook client allow MS to train their AI on your emails? Even if you use GMail in Outlook MS has access to it?
Any app beyond the absolute basics uses the (yes, sandboxed) Google Play store. But almost every Google play store app then uses Google Play Services to read your location etc., so you have to give that data to Google's software first if you want any app to work.
To get around it every app (or at least the ones like Lyft/Uber and airline apps that you need to move around) will need to be rewritten to avoid Google Play Services.
Not true.
All that's needed is for open source developers to "re-implement Google’s proprietary Android user space apps and libraries".
In most cases, the apps themselves won't know or care if they are feeding data to Google or not. For those few that do care, you can usually find a suitable replacement that doesn't --- or just use the publisher's good old-fashioned web site (aka banks) with a privacy focused browser like Brave.
Privacy invasion is baked into everything they do and only increases over time. The objective being to use your privacy to enrich themselves and bilk advertisers at your expense.
Great, so now we won’t even be able to have a choice around having our private messages off their AI surveillance machine for the Five Eyes… because just the mere sending a message to someone who has an Android could be doing this.
Buckle up folks, one day the Trump Junta or their descendants will be using Google’s SurveilAI tech to come for all you true Trump haters (incl me!)
Bullies like that with power are like rabid dogs.
I bet they crawl the Google Cloud block and object storage volumes and extract information too, which means that all SaaS hosted at Google is probably mined by Google. That would be my expectation unless the contract contains very specific non-weasel-worded language prohibiting this. Any company that does ads or ad-tech as a significant part of their revenue is guilty of maximal privacy invasion unless proven innocent.