Google banned from listening to Google Home recordings for three months in EU | Hacker News Reader

Google banned from listening to Google Home recordings for three months in EU

techerati.com

44 pointsolivegarten8 comments

I feel like this part of the article is a bit misleading:

> it emerged the company had been listening to intimate conversations of customers, even when users hadn’t activated the device by saying the phrase “Ok Google”. [ ... ] In a blog post, Google later admitted the claims were accurate.

While not outright stating it, this makes it sound as if Google admitted that listening regardless of whether the user said "OK Google".

Here's what the blog post (https://www.blog.google/products/assistant/more-information-...) actually says:

> The Google Assistant only sends audio to Google after your device detects that you’re interacting with the Assistant—for example, by saying “Hey Google” [ ... ] Rarely, devices that have the Google Assistant built in may experience what we call a “false accept.”

>While not outright stating it, this makes it sound as if Google admitted that listening regardless of whether the user said "OK Google".

What exactly is misleading? Sure they claim they weren't doing it intentionally [1], but it doesn't change the fact that there were instances of Google listening when the user hadn't said the phrase. Presumably the agency doesn't find that an adequate excuse.

[1] The sentence right after where you stopped quoting at:

>This means that there was some noise or words in the background that our software interpreted to be the hotword (like “Ok Google”).

I have no love for Google (nor should anyone.. :) ), but I agree with the parent that it was a bit misleading because the implication from that statement is that google just turns on the mic and records everything.

Yeah, that is a bit of a stretch, but still, there are a number of problems here. Google trumpets its artificial "intelligence" capabilities as they always do, and it eventually comes out that that its not "just a harmless computer algorithm" thats looking at your email or photos or whatnot, they give access to humans too. Oh, but only in a small tiny negligible, really just forget about this tiny tiny number, of cases.

"[...] confirmed that he worked for a sub-contractor, paid by Google, to transcribe and annotate recordings, including presumptions of the speaker's age and other demographics."[1]

We don't collect any personal data. Its just non-personal stuff to feed into our algorithm. Promise!

Also Google: "We partner with language experts around the world to improve speech technology by transcribing a small set of queries - this work is critical to developing technology that powers products like the Google Assistant"[1]

Seriously? What kind of BS logic is that. Let me just snoop on everyone so that we can improve our products to make more money, after we just made money selling you the product. They should be paying people to sign up for a study where they put in a listening device in their home to train their algorithm. Or purchase voice data from a corpus. Now they can ethically listen in on whatever they want and whenever they want. Instead, they just bury this under some opaque lawyerly language and spy on everyone. Is this the same company that purports to employ the smartest genius level talent?

[1] https://www.theinquirer.net/inquirer/news/3078773/google-ass...

> Let me just snoop on everyone so that we can improve our products to make more money ...

MS uses exactly the same twisted logic for their telemetry collection. :/

Agreed, Any telemetry that touches personal documents or data should be an opt-in.

I was given a google home device for free and actually, I doubt that there are "false accepts". I mean with alexa you can just say "alexa, do ..." with google home you basically need to scream and speak correctly that it turns it lights on.

> I was given a google home device for free and actually, I doubt that there are "false accepts". I mean with alexa you can just say "alexa, do ..." with google home you basically need to scream and speak correctly that it turns it lights on.

I've seen it happen a few times with the TV on or multiple people talking in the room. It's rare, but it does happen.

It's still far more innocent than the article suggests. There is usually a very clear indicator (lighting up on the top) of when it hears the trigger phrase. That paired with the low frequency of false accepts makes it a non issue.

Reply on news.ycombinator.com