This should be standard info at this point if you buy one of these things, but I still find Amazon's specifically-crafted responses to this so frustrating.
Out of one side of their mouth is the standard obfuscation with not selling _personal information_ or _requests_ per se:
> "We are not in the business of selling our customers' personal information and we do not share Alexa requests with advertising networks."
But out of the other side is indeed confirmation that they're sharing the "results":
> "Developers get the information necessary to fulfill your requests within their skills, such as answers when you play a trivia skill, or the name of the song you want to play..."
Maybe they're not actually sending raw audio or transcripts (or they are in some cases or with some partners, I would not be surprised about that either), but they're sending some amorphous amount of "information necessary to fulfill [our] requests" with no specifics obviously. A non-shady company would clearly articulate what they're sharing.
How do you trust this at all, or not just assume it's all going to the "41 advertising partners" they currently share it with?
> Amazon confirmed to The Verge that it does use voice data from Alexa interactions to inform relevant ads shown on Amazon or other sites where Amazon places ads. “Similar to what you’d experience if you made a purchase on Amazon.com or requested a song through Amazon Music, if you ask Alexa to order paper towels or to play a song on Amazon Music, the record of that purchase or song play may inform relevant ads shown on Amazon or other sites where Amazon places ads.” Amazon spokesperson Lauren Raemhild said in an email.
> Amazon told The Verge that it believes the research is flawed. “Many of the conclusions in this research are based on inaccurate inferences or speculation by the authors, and do not accurately reflect how Alexa works,” Raemhild said. “We are not in the business of selling our customers’ personal information and we do not share Alexa requests with advertising networks.”
So if you make a purchase via Alexa it informs your recommendations like if you made a purchase via the website? If this is the case, then the title is incorrect: the voice data is not being used, it's the actual purchase history that's being used.
All this talk about smartphones and Alexa listening to conversations to power ads fail to comprehend the economics of it. Constantly recording and analyzing audio to parse out relevant conversations would take lots of compute power and transmission usage. I am incredibly skeptical that this would yield enough of an advertising edge to make it a net positive. And this isn't even touching on the reputation hit of companies did this.
Yes! It's so common for people to be absolutely convinced that their phone or some app on their phone is listening to them in order to know what ads to show but it just doesn't make sense:
1. The ad company wants to target you. If your friend is talking the voice interpreter needs to be able to determine that it is not you talking. More than that, it would potentially need to be able to differentiate your voice from a multitude of voices at some social function.
2. Once your voice has been isolated then we're talking about natural language processing.
Admittedly, the technology required to do this is very close if it's not already here (GPT-3, etc.) but, to your point, it would at least not be a good look if Amazon (or Facebook, or Google, etc.) did this.
> What reputation hit? The past 20 years of tech has shown that people in general don't care about privacy.
I don't think reputation is really an issue for most major companies at all. People can't keep up on every sin committed by corporations and quickly forget.
Gerber and Beech-Nut have been found to sell baby food full of heavy metals. Literal poisons in dangerous amounts spoon fed into the mouths of our children, but they haven't gone out of business. DuPont knowingly poisoned people (and ultimately every last one of us) but somehow that doesn't matter. They're still operating and profiting. There are countless examples of companies who have done plainly evil things leading to deaths and suffering and yet they face zero meaningful consequences. Not from government, and not from people "voting with their wallet"
Companies only seem to care about their reputation because it can have an impact on profits in the short term which happens to be what many companies are overly focused on anyway. Don't want stock prices to drop even a little for a few months before everyone moves on to the next scandal.
Even if people don't care about privacy, they'll care about the network costs and battery drain from transmitting all this audio. If the audio is being parsed on the edge device it'd cut down on network costs, but result in even more battery drain.
Out of one side of their mouth is the standard obfuscation with not selling _personal information_ or _requests_ per se:
> "We are not in the business of selling our customers' personal information and we do not share Alexa requests with advertising networks."
But out of the other side is indeed confirmation that they're sharing the "results":
> "Developers get the information necessary to fulfill your requests within their skills, such as answers when you play a trivia skill, or the name of the song you want to play..."
Maybe they're not actually sending raw audio or transcripts (or they are in some cases or with some partners, I would not be surprised about that either), but they're sending some amorphous amount of "information necessary to fulfill [our] requests" with no specifics obviously. A non-shady company would clearly articulate what they're sharing.
How do you trust this at all, or not just assume it's all going to the "41 advertising partners" they currently share it with?