I mean if they were doing this already there would be no point in sending this email out. They would have just happily continued letting people think it wasn’t happening while doing it anyway, while not having to deal with the backlash this will generate.
My suspicion is they probably need to announce it now for some legal reason but there's no Amazon device with the power to do this locally so it's definitely always been sent to them.
Now would they delete that right away or analyse it first, I kinda think they would have always done the latter.
For anyone with existing Home Assistant setup, the Home Assistant Voice Preview is pretty good alternative, when it comes to voice control of HA. The setup is very easy. If you want conversational functionality, you could even hook it up to an LLM, cloud or local. It can also be used for media playback and it's got an aux out port.
I used to use Google Home Mini for voice control of Home Assistant. The Voice Preview replaced that rather nicely.
I'm using a Pi Zero as a voice satellite with an additional mic hat and a speaker hanging off the audio output and it's ... ok
There's definitely much lower WAF with this option
The voice assistant has built-in audio which appears to be high (enough) quality and considering it's case, power, etc, not to mention funding the advancement of open source voice control, it's just overall “better”
If you've got a Pi lying around with a mic & speaker, definitely give it a go
I agree. Although it’s nearly impossible at this point. Especially with Amazon running a significant portion of the internet with AWS. Each one of us most likely touches an Amazon server multiple times a day, even if we don’t have any Amazon subscriptions.
That doesn't matter. You only need to worry about boycotting things within your control, like Amazon shopping and their consumer products. AWS is profitable, but so is Amazon.com.
Buying something at a different store is always a dub even if that store is using AWS on the backend.
Like the other person said, you can at least control what you interact with directly. So you cancel your Prime subscription and turn your lights with your hand instead of an Echo but you don’t worry so much about trying to figure out if any of the several companies involved in making [product] have some form of attachment to AWS.
And there will be some level of consumption in this horrible system that’s not gunna be good in order for you to not be horribly depressed but people can shed more than they think and alternatives do exist for many of the ones you might put at lower priority.
If anyone else is wondering, I’ve not found a verbatim quote of the steps but I did see an article that mentioned the consequences. It seems like you will be able to turn this off but it will disable Voice ID:
anyone with their Echo device set to “Don’t save recordings” will see their already-purchased devices’ Voice ID feature bricked. Voice ID enables Alexa to do things like share user-specified calendar events, reminders, music, and more. Previously, Amazon has said that "if you choose not to save any voice recordings, Voice ID may not work." As of March 28, broken Voice ID is a guarantee for people who don't let Amazon store their voice recordings.
The old "privacy focused" setting made speech processing local. The new "privacy focused setting" means that processing will happen on a remote server, but Amazon won't store the audio after it's been processed. Amazon could still fingerprint voices with the new setting, to know if it was you or your parents/parter/kid/roommate/whomever and give a person specific response, but for now at least they appear to not be doing so.
This all seems like it's missing the point to me. If you own one of these devices you're giving up privacy for convenience. With the old privacy setting you were still sending your processed speech to a server nearly every time you interacted with one of those devices because they can't always react/provide a response on their own. Other than trying to avoid voice fingerprinting, it doesn't seem like the old setting would gain you much privacy. They still know the device associated to the interaction, know where the device is located, which accounts it's associated with, what the interaction was, etc. They can then fuse this information with tons of other data collected from different devices, like a phone or computer. They don't need your unprocessed speech to know way too much about you.
@richardisaguy@Tea sometimes they just come free with stuff. We got given two Google ones when my husband bought a Pixel phone. We were going to sell them on but we never got round to it. You can physically turn off the microphone part though (at least it tells you it's turned off so fingers crossed) so we use the one with a screen as a digital photo frame (and a speaker) and the other one as just a speaker.
True, but a mobile phone is basically a world brain, calculator, camera, flashlight, you can watch movies on it in hi def, hate it all you want, it's one of the most versatile tools on the planet.
An echo dot, it just spy garbage and nothing else
At least, on mobile devices, it's typically easier to install a privacy-focused firmware (like LineageOS or GrapheneOS). Those AI assistants are completely locked down.
Plenty of people I know have gotten the little echo dots or the bigger alternative with larger speakers for Christmas or birthdays. Technically they didn't spend money, but their friends and family did.
They literally could just leave the feature on the device, but then you can't force your users to send you all their data, voices, thoughts and first borns
Just sold my 3 devices and shut down Amazon account. It's very liberating and I don't miss it one bit. Have Home Assistant and a couple of really good 2nd hand Sonos speakers.
Have a look at Home Assistant! It’s a great open source smart home platform that recently released a local (so not processing requests in the cloud) voice assistant. It’s pretty neat!
home assistant is amazing but it is not yet an alternative to Alexa, the assistant/voice is still in development and far from being usable. it’s impossible for me to remember the specific wording assist demands and voice to text is incorrect like nine out of ten times. And this includes giving up on terrible locally hosted models trying out their cloud which obviously is a huge privacy hole, but even then it was slow and inaccurate. It’s a mystery to me how the foss community is so behind on voice, Siri and Google Assistant started working offline years ago, and they work straight on a mobile device.
I have one big frustration with that: Your voice input has to be understood PERFECTLY by TTS.
If you have a "To Do" list, and speak "Add cooking to my To Do list", it will do it! But if the TTS system understood:
Todo
To-do
to do
ToDo
To-Do
...
The system will say it couldn't find that list. Same for the names of your lights, asking for the time,..... and you have very little control over this.
HA Voice Assistant either needs to find a PERFECT match, or you need to be running a full-blown LLM as the backend, which honestly works even worse in many ways.
They recently added the option to use LLM as fallback only, but for most people's hardware, that means that a big chunk of requests take a suuuuuuuper long time to get a response.
I do not understand why there's no option to just use the most similar command upon an imperfect matching, through something like the Levenshtein Distance.
They have doorbells to watch who comes to your house and when.
Indoor and outdoor security cameras to monitor when you go outside, for how long, and why.
They acquired roomba, which not only maps out your house, but they have little cameras in them as well, another angle to monitor you through your house in more personal areas that indoor cameras might not see.
They have the Alexa products meant to record you at all times for their own use and intent.
Why do you think along with Amazon Prime subscriptions you get free cloud storage, free video streaming, free music? They are categorizing you in the most efficient and accurate way possible.
That is actually good news to hear. Not completely good on my part for being incorrect about ownership, but once I saw the proposed deal back when it was announced, I immediately added them to the “no I don’t think I will.” list of products I won’t support.
If anyone remembers the Mycroft Mark II Voice Assistant Kickstarter and was disappointed when development challenges and patent trolls caused the company's untimely demise, know that hope is not lost for a FOSS/OSHW voice assistant insulated from Big Tech..
Disclaimer: I do not represent any of these organizations in any way; I just believe in their mission and wish them all the success in getting there by spreading the word.
Off-device processing has been the default from day one. The only thing changing is the removal for local processing on certain devices, likely because the new backing AI model will no longer be able to run on that hardware.
With on-device processing, they don’t need to send audio. They can just send the text, which is infinitely smaller and easier to encrypt as “telemetry”. They’ve probably got logs of conversations in every Alexa household.
If you look at the article, it was only ever possible to do local processing with certain devices and only in English. I assume that those are the ones with enough compute capacity to do local processing, which probably made them cost more, and that the hardware probably isn't capable of running whatever models Amazon's running remotely.
I think that there's a broader problem than Amazon and voice recognition for people who want self-hosted stuff. That is, throwing loads of parallel hardware at something isn't cheap. It's worse if you stick it on every device. Companies --- even aside from not wanting someone to pirate their model running on the device --- are going to have a hard time selling devices with big, costly, power-hungry parallel compute processors.
What they can take advantage of is that for a lot of tasks, the compute demand is only intermittent. So if you buy a parallel compute card, the cost can be spread over many users.
I have a fancy GPU that I got to run LLM stuff that ran about $1000. Say I'm doing AI image generation with it 3% of the time. It'd be possible to do that compute on a shared system off in the Internet, and my actual hardware costs would be about $33. That's a heckofa big improvement.
And the situation that they're dealing with is even larger, since there might be multiple devices in a household that want to do parallel-compute-requiring tasks. So now you're talking about maybe $1k in hardware for each of them, not to mention the supporting hardware like a beefy power supply.
This isn't specific to Amazon. Like, this is true of all devices that want to take advantage of heavyweight parallel compute.
I think that one thing that it might be worth considering for the self-hosted world is the creation of a hardened network parallel compute node that exposes its services over the network. So, in a scenario like that, you would have one (well, or more, but could just have one) device that provides generic parallel compute services. Then your smaller, weaker, lower-power devices --- phones, Alexa-type speakers, whatever --- make use of it over your network, using a generic API. There are some issues that come with this. It needs to be hardened, can't leak information from one device to another. Some tasks require storing a lot of state --- like, AI image generation requires uploading a large model, and you want to cache that. If you have, say, two parallel compute cards/servers, you want to use them intelligently, keep the model loaded on one of them insofar as is reasonable, to avoid needing to reload it. Some devices are very latency-sensitive --- like voice recognition --- and some, like image generation, are amenable to batch use, so some kind of priority system is probably warranted. So there are some technical problems to solve.
But otherwise, the only real option for heavy parallel compute is going to be sending your data out to the cloud. And even if you don't care about the privacy implications or the possibility of a company going under, as I saw some home automation person once point out, you don't want your light switches to stop working just because your Internet connection is out.
Having per-household self-hosted parallel compute on one node is still probably more-costly than sharing parallel compute among users. But it's cheaper than putting parallel compute on every device.
Linux has some highly-isolated computing environments like seccomp that might be appropriate for implementing the compute portion of such a server, though I don't know whether it's too-restrictive to permit running parallel compute tasks.
In such a scenario, you'd have a "household parallel compute server", in much the way that one might have a "household music player" hooked up to a house-wide speaker system running something like mpd or a "household media server" providing storage of media, or suchlike.
People are saying don't get an echo but this is the tip of an iceberg. My coworkers' cell phones are eavesdropping. My neighbors doorbells record every time I leave the house. Almost every new vehicle mines us for data. We can avoid some of the problem but we cannot avoid it all. We need a bigger, more aggressive solution if we are going to have a solution at all.
I didn't even know this was a feature. My understanding has always been that Echo devices work as follows.
Store a constant small buffer of the past few seconds of audio
Locally listen for the wake word (typically "Alexa") using onboard hardware. (This is why you cannot use arbitrary wake words.)
Upon hearing the wake word, send the buffer from step one along with any fresh audio to the cloud to process what was said.
Act on what was said. (Turn lights on or off, play Spotify, etc.)
Unless they made some that were able to do step 3 locally entirely I don't see this as a big deal. They still have to do step 4 remotely.
Also, while they may be "always recording" they don't transmit everything. It's only so if you say "Alexaturnthelightsoff" really fast it has a better chance of getting the full sentence.
I'm not trying to defend Amazon, and I don't necessarily think this is great news or anything, but it doesn't seem like too too big of a deal unless they made a lot of devices that could parse all speech locally and I didn't know.
Which Echo devices ever supported local only processing? They cost about £30. There's no kit that can do decent voice commands for that money. You'd be lucky to have a device that processes claps to turn the lights on for that.
If you traveled back in time and told J. Edgar Hoover that in the future, the American public voluntarily wire-tapped themselves, he would cream his frilly pink panties.
be aware, everything you say around amazon, apple, alphabet, meta, and any other corporate trash products are being sold, trained on, and sent to your local alphabet agency. it's been this way for a while, but this is a nice reminder to know when to speak and when to listen
How disheartening. I knew going in that there would be privacy issues but I figured for the service it was fine. I also figure my phone is always listening anyway.
As someone with limited mobility, my echo has been really nice to control my smart devices like lights and TV with just my voice.
Are there good alternatives or should I just accept things as they are?
There aren't any immediate drop in replacements that won't require some work, but there is Home Assistant Voice - It just requires that you also have a Home Assistant server setup, which is the more labor intensive part. It's not hard, just a lot to learn.
Thanks for the tip. I had thought of doing Home Assistant to automate things like lights but I use the Echo for other stuff too (like reminders and weather). I guess I'll just wait for the open source tech to catch up.
And for now it's voice assist is garbage in comparison. I have home assistant, and a few Alexa units, so I set up nabu and tried it, but it's slow and can maybe do 1 in 5 commands, while Alexa is much more reliable.
Easy fix: don't buy this garbage to begin with. It's terrible for the environment, terrible for your privacy, of dubious value to begin with.
If every man is an onion, one of my deeper layers is crumudgeon. So take that into account when I say fuck all portable speakers. I'm so tired of hearing everyone's shitty noise. Just fucking everywhere. It takes one person feeling entitled to blast the shittiest music available to ruin everyone in a 500yd radius's day. If this is you, I hope you stub your toe on every coffee table, hit your head on every door jam, miss every bus.
Today: "...they will be deleted after Alexa processes your requests."
Some point in the not-so-distant future: "We are reaching out to let you know that your voice recordings will no longer be deleted. As we continue to expand Alexa's capabilities, we have decided to no longer support this feature."
And finally "We are reaching out to let you know Alexa key phrase based activation will no longer be supported. For better personalization, Alexa will always process audio in background. Don't worry, your audio is safe with us, we highly care about your privacy."
So... if you own an inexpensive Alexa device, it just doesn't have the horsepower to process your requests on-device. Your basic $35 device is just a microphone and a wifi streamer (ok, it also handles buttons and fun LED light effects). The Alexa device SDK can run on a $5 ESP-32. That's how little it needs to work on-site.
Everything you say is getting sent to the cloud where it is NLP processed, parsed, then turned into command intents and matched against the devices and services you've installed. It does a match against the phrase 'slots' and returns results which are then turned into voice and played back on the speaker.
With the new LLM-based Alexa+ services, it's all on the cloud. Very little of the processing can happen on-device. If you want to use the service, don't be surprised the voice commands end up on the cloud. In most cases, it already was.
If you don't like it, look into Home Assistant. But last I checked, to keep everything local and not too laggy, you'll need a super beefy (expensive) local home server. Otherwise, it's shipping your audio bits out to the cloud as well. There's no free lunch.
I honestly have no idea why anyone who cares even 1% about their privacy would have ever bought one of these abominations in the first place. If I ever receive one as a gift I will burn it with fire.
I have the things so that I can understand how to protect myself from them. I have a similar thing going on with AI video right now. Hate it but watch the growth to understand it.
Me while cooking mac and cheese for the kids:
"Echo, set timer for 8 minutes"
Echo: "GOOD EVENING [me], SETTING TIMER FOR 8 MINUTES"
No, shut the fuck up and just set the goddamn timer without the extra fluff. I've seen Ex Machina, I know you have no empathy, so knock off the "nice" shit and do what I fucking ask without anything else.
These devices are a great help to the elderly and disabled, and Amazon's offerings are the least expensive. My buddies Mom can call emergency services from anywhere in her home with just her voice. I don't think that's foolish at all.
The part that really gets me is that you have to opt out to not have everything you say saved. Bonkers that that isn't the default! There's no good user-based reason for this. Alexa doesn't remember shit for users, like any AI there's no recall feature. You can't say remember what I told you last night - give the address for that place, I was drunk and don't remember the name.
be aware, everything you say around amazon, apple, alphabet, meta, and any other corporate trash products are being sold, trained on, and sent to your local alphabet agency. it's been this way for a while, but this is a nice reminder to know when to speak and when to listen