4 Boosts
0 Likes
This comment is unavailable — it may have been deleted or is not visible to you.
This comment is unavailable — it may have been deleted or is not visible to you.
This comment is unavailable — it may have been deleted or is not visible to you.
Not in the least better for privacy to process audio into text and delete the audio. It just reduces the bandwidth required to exfil.
Just learned that Smart TVs also process audio to test within 70 feet radius and send text to Corporate for analysis
@black @M0PWX @LeslieBurns Not really seeing anything particularly unusual in this case. As long as standard wake word and/or remote activation is used for voice activation, with local processing prior to that, it's in the standard model for voice devices (e.g. Google Home, Alexa, etc.) Network scanning is also completely typical for IoT devices. Of course dedicated hacking can often disrupt things, especially by local attackers on the same network, but remote hacks of smart TVs are indeed much rarer. Overall, from what I've seen, this is pretty much a typical "smart TV" with the usual pluses and minuses.
@lauren @black @M0PWX @LeslieBurns I'm not sure that many people realise quite how bad a "typical smart TV" really is in terms of the information it gathers, retains and forwards from what most would regard as a private space.
@rvkennedy @black @M0PWX @LeslieBurns It's usually not awful as long as ACR is turned off. Even if you use a dumb TV (or just don't hook a smart TV up to the Net and only use HDMI inputs) the streaming services, etc. are going to know what you're watching on them, by definition.
@lauren @black @M0PWX @LeslieBurns the gist of the GN report on LG though was that the TV never stops listening, even on standby. It scans your local network to identify what devices you might own. And it uses not only what it knows you're watching (and ACR is itself a horrifying abuse of the users own resources for something that benefits them not at all) but also what it hears you saying. At least that's my understanding of the documentary, it's pretty long and I zoned out after 1/2 hour or so
@rvkennedy @black @M0PWX @LeslieBurns Voice devices are typically always listening because they're waiting for the wake word. So long as the processing for that is on device, and data doesn't get sent out until AFTER the wake word is detected, it's S.O.P.
@lauren @black @M0PWX @LeslieBurns that's not the behaviour GN describes in their show...
@rvkennedy @black @M0PWX @LeslieBurns The only aspect that really matters is out of scope data (ACR info if ACR is disabled, ambient voice data, etc.) actually being SENT from the device to an external server. Local capture capabilities are really not an issue. Saying that the device COULD do this or COULD do that is not the issue. It's what it's actually doing that matters. This applies to every connected computing device essentially everywhere.
@rvkennedy @black @M0PWX @LeslieBurns So for example the new Apple watch is a privacy problem (in 2-party consent states, EU, etc.) -- I'll be talking about this on my regular national radio segment Monday --- because it actually IS sending data related to an eavesdropped conversation to Apple servers, without permission or notification to all parties.
@lauren Not legal in California.
Gotta be kidding…
@lauren but what how surprising
@lauren Seriously? US only I guess? I don't see how that could possibly be legal in Europe.
@MatthewToadAgain Dunno yet about the geo aspects outside U.S.
@lauren @MatthewToadAgain Yeah I'd imagine that's very much illegal in the UK and Europe. Hope so anyway!
Is it an option that can be disabled?
@Defaulty The user enables it. The problem is that anyone around them sucked into it has no choice at all and no notification from the watch.
