@mttaggart
> looking at this settlement and ascribing guilt in this case is prejudicial
Aww, poor Goggle. Why don't people trust them? It's so unfair. Honestly, is this really a hill worth dying on?
What's hypothetically true - and again we don't know - is much less important than how we act on the possibilities.
@mttaggart people failing to understand that settling a lawsuit is not the same thing as losing in court will never cease to disappoint me. But I guess it lets them draw the conclusion they want regardless of reality. FUD peddling.
@mttaggart
thats a lot of money thrown at "nothing to see here"
@TheDailyBurble It is under 20 minutes' worth of revenue for them.
@mttaggart Please don't dump on me...
I've personally observed both Alexa & Siri activate without the trigger phrase being uttered.
I knew because I heard the audio responses they both gave. Had the volume been down?? In Siri's case, (this past weekend) it called a contact whilst I & a friend were talking. No trigger phrase or anything close was uttered.
So, if this stuff is buggy, they could absolutely be recording conversations, and be thinking with consent. But also, no obvious trigger..
@mttaggart I kinda hope they ARE listening. Because its scarier if *they don't have to be listening*.
If -- Yes you are, in fact, just that predictable. If, Between your browsing habits, social media intake, video feed and the same data on your connected social graph, they can predict very well what you want.
Well that is some "Devs" (great show, highly recommend if you like mind bending SciFi) level existential crises around the notion of free will. And that idea is just too scary for some people. (Me included. I just try to ignore it)
@varx They built the maze and they know all the corners you'll run to and why.
@mttaggart i went through wife's phone and deleted the saved audio files of conversations google recorded while my wife just had her phone in her pocket.
@mttaggart I've always suspected its confirmation bias, driven by other surveillance techniques.
Did you talk to somebody about something, and suddenly get a bunch of ads for it? Did that person recently google it, shop for it, engage somehow online? Well, location tracking knew you two met up, so... advertise the stuff your fiends are interested in
It is not at all trivial to activate a microphone without user knowledge, and reverse engineers have been looking for a smoking gun to prove Google or Apple or Meta do this for years with nothing to show for it.
The "proof" most people have is the result of very good ad targeting, not eavesdropping.
@mttaggart I agree with you it is probably a bullshit claim (and companies like google can easily buy it off, see https://infosec.exchange/@avuko/116000724082095895).
But there is the uncomfortable truth that any device which responds to a "trigger word", needs to listen 24/7 for that trigger word. That means a microphone is always-on, even if just listening in small loops going to /dev/null.
My guess was always that the processing of the audio, to find the trigger word or for other analysis, is on-device.
Otherwise you probably couldn't do a "hey google" (or similar for other brands) in airplane/offline mode.
I don't have an android phone, nor any of these services enabled, but I guess they work in offline mode?
@avuko Right although that's a little different than audio capture and send, which is the claim here.
@mttaggart agreed. I personally never bought the "they are listening to sell ads" stuff. They have all of your online behaviour, plus everyone in your graph, including the people you app with, and what they are looking at while talking to you, clicking on, sending you etc. etc. ad nauseam.
Why would they need to spend time and money and effort to process what you say?
And! But also completely unrelated! (gestures wildly)
Yes, there is a corpo-owned black-box device with an always-on microphone in the room. Often multiple ones. It also has fine grained location tracking, and has you logged into a gazillion tracking services. We call that a mobile phone.
For the love of Marduk I am not "defending" Google. I am asking that claims about your adversary's (and giant tech companies are adversaries) capabilities have evidence to back them up.
@mttaggart inb4 you get a bunch of "evidence" that's actually just links to conspiracy theory YouTube channels.
@cR0w Can't wait. Or for Gemini/IPFS links
@cR0w It's still with us, and guess what else it inspired? https://atproto.com/specs/data-model#relationship-with-ipld
@mttaggart It also just seems like a really bad way to get ad data. I'm usually not sitting there yelling, clearly, about how I need new socks. The million other sources of adtech surveillance data is already plenty (and yet they still haven't figured out I need new socks).
@mttaggart It would also be plain as day in a pcaps.. you can monitor everything and wouldn't see it.
I'm very curious what's going on, and hate that this is confirming a huge conspiracy theory
@iagox86 I mean I think what's "going on" here is that it cost Google very little to settle, and so they did. No need to go through discovery and potentially a much costlier set of penalties if the case didn't go their way.
And then you have an entire subset of social media that makes their hey on ascribing guilt based on reputation and speculation.
Part Baader-Meinhof phenomenon, part ad targeting.
Even if you weren't the one to search for the thing, if a person who texts you a lot or whose location is near you or connected to your WiFi then might as well spread some targeted ads to the people who might have talked to one another about it.
You probably get ads for the people around you all the time and most of the time they're just irrelevant whatever. But sometimes they do hit right
@mttaggart I ran my Echo on a MITMed lan for weeks and nada.
@mttaggart
> Google has any number of reasons to not want to go to court on this
Really? They've got deep pockets and laywers on speed dial. If it really isn't possible that ...
> a microphone is being activated by a Google service without user input
... surely the best way to put paid to the speculation is to put it before the courts? Settling out of courst doesn't prove it's happening, but they've missed a prime opportunity to convince people it's not, so ... 🤷♂️
@strypey Discovery is never a preferred process for a company like Google when settlement is this cheap. It proves nothing. As I said, inferring guilt based on settlement is speculation.
As I mentioned above, great reverse engineers have been looking for this for a while and found nothing. I really don't think this is a thing the way most seem to assume it is.
@mttaggart
> great reverse engineers have been looking for this for a while and found nothing
Absence of evidence is not evidence of absence, and they have pretty good reason to hide this stuff *very* well. Knowing that reverse engineers will be looking for it, and that the consequences of them finding it would be *very* bad for the company.
> I really don't think this is a thing the way most seem to assume it is
Lots of people report experiences that can't be explained otherwise.
@strypey They can be explained otherwise! It's good ad targeting! And "absence of evidence" only holds up when you don't have the means to collect the data. We do! We have! Those things simply are not present in the apps and services after heavy scrutiny.
(1/2)
@mttaggart I guess it comes down to what makes sense as a default assumption in the face of a known unknown. Given everything we know about Goggle, it is a safer assumption that they designed ChromeOS/ Android to give them access to mics, or that they haven't?
Combine that with a variant of Pascal's Wager; what are the worst consequences of being wrong about the first assumption, vs. the second one?
(2/2)
For example, what If I'm an anti-ICE protester with an Android device, and I wrongly assume Goggle can access my mic? Then I've been slightly more careful than I needed to. Not so bad.
But what if I wrongly assume Goggle *can't* access my mic, and they record my protest planning discussions, and they're passed to ICE via the NSA dragnet Snowden exposed? Potentially *very* bad, and not just for me.
@strypey That is a moved goalpost. The discussion was never about what to give Google, but whether they are doing what is claimed. And there is no evidence they are eavesdropping. So looking at this settlement and ascribing guilt in this case is prejudicial.
![unlofl [Promoted Toot]](https://media.mstdn.social/accounts/avatars/109/139/379/925/966/320/original/e28a291a31f5844b.png "unlofl [Promoted Toot]")
