Spotify created a patent. Does that mean they will actually use it? Plenty of companies file for patents for all kinds of ideas, just to have a patent in the warchest for litigation and to hold IP.
Willing to bet that you're using at least one piece of software on a daily basis that's made by a company that - just like Spotify - files "ridiculous" patents all the time.
The issue with replacing Spotify with a self hosted alternative is the need for content.
Spotify (EDIT: for some people) solves the content problem for a low fee. Self hosted requires either spending lots of money, or in lots of regions something that probably qualifies as copyright infringement.
(EDIT: noted that this isn't universally solved, and thanks for responses pointing this out)
I can understand that. In many ways people become rely on this big Internet "Database" to hear and listen to things. However, companies runs Internet become more censored and surveilled. Disk and USB space are cheap, it is a good idea to start archiving and collecting things where you can play without Internet.
> > Self hosted requires either spending lots of money, or in lots of regions something that probably qualifies as copyright infringement.
> Disk and USB space are cheap
Apologies for any ambiguity. When I was refering to spending lots of money, I wasn't considering the costs of self hosted infrastructure (servers, backups), but rather the cost of acquiring the media in the first place.
This is kind of an overreaction. Companies file patents for all kinds of things all the time, most of which never become real products. And if this one did, I can just deny Spotify access to my microphone on iOS and macOS. Not a big deal.
I wonder if TikTok and Instagram are already doing this. Specially with cheap on device ML systems you can easily extract emotions from spoken words in Instagram stories or TikTok videos
Unlikely this patent will ever get used for something, but I think it could actually be used for good: if there's some music that makes me happy and some music that makes me sad without realizing it, it would be nice for "Spotify" to detect this and let me know/change the style of the music.
Facebook some years ago already played with people's emotional well-being in tests. A lot of the direct links to the articles have somehow been scrubbed from the internet, but here's the HN results: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
Unbridled data collection is a requirement for a startup? Maybe in the US, pump-and-dump-startup schemes; I like to think other countries found companies that actually do something useful.
I mean Spotify has been one of the biggest driving forces in reducing music piracy, in part thanks to being more convenient.
On the other hand, they don't pay artists nearly enough so the flipside is that to the artists it doesn't actually make that much of a difference whether someone listens to them on Spotify or just pirates the music.
Really wonderful, sustainable businesses when the inevitable privacy legislation hits the books in the United States. Just great high quality revenue, wish I could get in on some of that
I understood the point until the site stated "Emotion recognition software is largely seen as racist pseudoscience" and then link to facial recognition which is not related.
You'd think these people are smart enough to not overplay their hand and actually get traction and therefor reach their goal.
I didn't think so. The second sentence explains the petition is a call to "abandon this patent".
I don't think the author knows what a patent is, because I also don't get the impression that the point of this is make the idea public domain instead of owned by one company.
Or speaking of paranoia (you weren't but I did in my other comment) maybe they DO want it to be public domain, because maybe this whole anti-Spotify campaign is created by Facebook, because Facebook doesn't want a Spotify patent to prevent Facebook from doing this technology themselves.
yeah, that bothered me too. If it's a pseudoscience then why should I care. I hate it when such campaigns with very concrete objectives try to drag everything possible into them.
This inevitably muddies the water and leads to semantic overload, where everything is racist.
Muddying the water is phrasing it well.
Don't get me wrong. I think pseudoscience and racism etc. are all bad things that need to be solved as well.
However adding them on one big pile does not make your cause better or make the problem go away.
Actually it scares people like me away, because it shows you're on a emotional slander campaign instead of solving a problem.
It's also a successful tactic, sadly, to just call everyone a racist, the worst thing you can call someone nowadays, and hope they cave. I'm simply not having it and will never engage in something like that.
Honestly, it feels a bit 'counter-intelligence'-y to me. Like, it's so over the top that I almost suspect Spotify did it themselves.
The Whois on the site says that it was registered on April 5th of 2021 for 1 year and the customer is based out of Ontario, Canada. Tucows was the registrant, and they seem to be a generic handler of web domains, based out of Toronto. https://tucows.com/
Other than that, I can't really find out any additional info about the site or who is running it.
I'm sure other HNers are more adept at searching for these kinds of things, and I'd love to know more about the paper-trail of it.
I feel like there is a ton referenced concerns in that first page I read that isn't really explained, some seem kinda 'off', and just makes me hit the back button.
I'm not singing up for what seems like a mix of legit concerns and ... borderline conspiracy theory / ideas that maybe really don't help the cause.
I'm inclined to sympathize with a lot of this but I'm not getting on board an unknown / the whole crazy train.
When I worked at IBM, I was explicitly told that I would be surprised with what I could patent, and that I should patent things even if they seemed trivial.
This went so against my ethos I quit and found another job.
The original iteration of Spotify scanned the local filesystem and shared content on the peer-to-peer network. That was enough to ban it from my work and home networks and personal devices. I am not interested in whatever Spotify is currently doing. They are not trustworthy as an organization.
It did those two separate things. It did not share your local files via P2P.
It scanned the filesystem to find local media files to present for playback. It used P2P to legally share locally cached parts of DRM:ed Spotify tracks to other users in order to lower Spotify's bandwidth costs.
It did indeed share local files. That was part of how they were able to handle the network load. If it found a file with a matching hash of something it had licensed, it shared it. Check out https://launch.co/blog/did-you-know-spotify-uses-peer-to-pee...
It was well known at the time. How do you know it was not sharing other files it found?
I was running fs_usage against the Spotify process in 2013. It was reading local files and sending data out to untrusted peers. It's okay that you think that is okay. I don't allow that on my machine.
That is not an extraordinary claim. Everybody knew that Spotify was scanning the filesystem. Some trusted that it was careful about what it uploaded. People with security concerns did not.
Spotify stopped doing P2P once it got big enough. I don't care what it is doing now. I already banished it from my devices and networks. Nobody needs pop music that badly.
Making use of of peer to peer protocols to distribute content is something I wish more companies did.
P2P lowers the barrier to entry for new players in a domain, because it allows us to bring a product to the consumers with potentially drastically lower cost of operating, because we can provide service to our users without the massive bandwidth costs to our business.
This is good for the consumer too. So don’t be mad at companies that use P2P technology. Be in favor of it. And demand better terms from your ISP instead. So that the internet can continue to allow for innovation and competition, instead of further centralizing all of the power to only the biggest companies and slowing down innovation and hurting competition.
I understand that it was good for the network. But the Spotify app was scanning directories other than its own (e.g., it scanned ~/Music on macOS looking in your iTunes library). It could have through negligence or malfeasance shared any file on the machine.
This shouldn't be up to Spotify – it should be illegal for apps to listen to people's conversations unless explicit to the user every time (eg; phone calls, voice memos), even if the contents of the conversation aren't sent off-device.
There is widespread suspicion that Facebook Inc already does this (they deny it, but I've heard and experienced too many anecdata to believe them).
Speaking as someone who can at times be a targeted-ads apologist, conversation-snooping software is an abhorrent invasion of privacy that I can't imagine anyone would be happy with.
Source? I haven't seen anything persuasive. And I'm not sure how this could be "easily testable" if they are being sneaky about it (using undocumented APIs in obfuscated ways, etc).
BTW, I didn't claim that they do this, only that I and others suspect they do.
Personally I've seen cases were a topic was discussed aloud which was not searched/browsed by either party and received ads very shortly thereafter. Extremely specific stuff, like you randomly talk about life insurance and then get a life insurance ad.
> Facebook and Facebook Messenger apps for iOS and Android devices consume a lot of battery life. Besides complaints from people worldwide, authorities and analysts conducted tests. They affirmed that both are battery hogs even when the apps aren't in use.
I'm not sure how one would be expected to notice a "large upload of audio" since most people don't monitor their network activity, and slowdowns get blamed on a bad connection. I'd also expect smart FB engineers would be able to figure out how to minimize network activity (eg; sampled, low-fidelity audio).
EDIT: To be clear, I don't mean to say, "they use lots of battery therefore they must be spying", just that lack of battery drain isn't convincing evidence that they're not doing some sort of conversational spying.
Well first and foremost, as an app on iOS/Android, you can't just access the mic whenever you feel like it. There's also an available audit trail for when it is accessed. So it's trivial to know whether the Facebook app uses the mic without being asked to do so by the user.
Unless you think they are 'being sneaky' by developing zero day exploits to break permissions boundaries and prevent logging on both Android and iOS.
What's more likely is that: (a) we're all a lot more predictable than we'd like to admit, and (b) you underestimate the multitude of data sources available to Facebook and piped over to its ecosystem. Just because an ad is on Facebook, doesn't mean Facebook is the one providing all the targeting data.
> Unless you think they are 'being sneaky' by developing zero day exploits to break permissions boundaries and prevent logging on both Android and iOS.
Pretty much, yes. Perhaps only one of the two operating systems, and perhaps not all versions.
I'm not confident they've done that, but it'd be hard to prove they haven't, and frankly it seems like the kind of thing FB might do (I'm sure not everybody would agree).
It's certainly possible that FB is just wildly superior at understanding/predicting people than humans are. Their targeting is, after all, really good.
> Queer indie-punk musician and Fight for the Future director Evan Greer has released a new music video for the song “Surveillance Capitalism,” ahead of the release of the album Spotify is Surveillance.
That raised a red flag for me. Why is this person's gender identity in any way relevant to the issue at hand?
In the same way the text calls out emotion recognition software as "racist pseudoscience" (then linking to an article about facial recognition) - which doesn't have much at all to do with the patent.
Bought to you by the "Union of Musicians" - the same people who claim Spotify make "enormous profits" (I wouldn't call them enormous) and want 1 cent per play.
They want Spotify to pay them, but they also want to damage Spotify's reputation? Talk about shooting yourself in the foot
Further to that, check out the Union of Musicians' subcommittees [1] - there's a "Classical Music Accountability" subcommittee (classical music = white supremacy?)... and a Police Abolition committee, of course.
There are definitely elements of racism rooted in Classical music. Classical music is not something that was historically accessible to the masses. Minorities have not always been allowed to participate in it the same way White people were allowed to. with a quick internet search you can find out more from more legitimate sources.
It's interesting because it's the first time I've thought about it (not a big listener of classical music) - but whether or not classical music is racist isn't really why I bought that one up.
The group seems to be sidetracked with every single cause, a subcommittee for every minority group... which, to me, takes away from the main purpose of a Union of Musicians - getting paid/treated properly (as a whole).
Which just reminds me of how this webpage gets sidetracked from the main issue. The main issue should be: listening in on conversations is creepy.
This trope is everywhere now. I see this so often now where high profile substack writers are constantly being smeared as transphobic. It's a tactic, and defense and offense strategy and I'm tired of it. Can't take anyone serious anymore how writes something like that.
Or you could be charitable to the author and at least try to understand their view instead of assuming that it’s some cold political move.
Machine learning detection of people’s gender in the real world is anywhere from annoying to distressing to trans people because all the training data is cis people and so they get matched with their AGAB. Can you imagine how terrible it would be to already hate and be super self-conscious about your voice only to have Spotify start telling you via recommendations that you don’t pass well enough. Yikes.
> Can you imagine how terrible it would be to already hate and be super self-conscious about your voice only to have Spotify start telling you via recommendations that you don’t pass well enough.
Although the patent may be creepy...
The text of the patent refers to media recommendation (not advertising) - contrary to the sibling comment where you claim it will result in ads for Dollar Shave Club, the patent does not once use the term "advertising". Sure, it could be used for that, but I don't think that's the main intention at all.
How are you going to know that Spotify has recommended you a song based on your gender? Are we assigning genders to songs now? Straight white cis males can like Aqua's "Barbie Girl", too.
> Are we assigning genders to songs now? Straight white cis males can like Aqua's "Barbie Girl", too.
This is an absurd take. Do you see no difference between Spotify recommending songs to a trans woman based on what men with similar music tastes in her area and age group are listening to and "assigning gender to songs?"
Spotify also typically recommends you playlists rather than songs and my home page is extremely gendered. I get playlists like "You Go Girl", "Girls Like You", "Women of Rock", "Fierce Femmes." A trans guy getting these kinds of recommendations would be ... well odd.
You don't need to be a woman to enjoy a "Women in Hip Hop" playlist. Spotify isn't about what it thinks you are. It's about what it thinks you want to listen to.
And what it thinks you want to listen to is based on what it gender it thinks you are. Gender is a strong signal about what your music tastes will be but obviously it's not absolute. If the algorithm is so pure and unbiased then why would it even want to determine your gender with this patent?
Like recommendation engines aren't magic -- they look at what people "like you" listen to where "like you" has all sorts of random demographic data as well as your personal patterns/history.
> It's about what it thinks you want to listen to.
At the moment, but I think the idea behind this patent is that it can factor in your mood (and other factors, such as gender) into the recommendation engine as well as what you already like.
Though I don't see gender being much of a signal. My GF and I have exactly the same taste in music (well, she doesn't like Classical that much, and I don't like Musicals, and she's more into Goth/Industrial/Metal than I am, but it's a 90% match). Anecdata, but that's all I have.
Spotify’s algorithms try to detect your gender as part of the profile they build about you. The training data will be all cis people and so trans people are likely going to get matched against their AGAB. This really sucks for trans people to have some system they have no influence over constantly serving them ads and music targeted at an identity that causes them distress.
A trans woman constantly getting ads for dollar shave club, Nivia for men, and Weezer is hugely invalidating simply because some machine learning algorithm says she sounds like a man. Good god imagine having to perform a stereotypical version of your gender identity for your phone just to get the recommendations you want.
Pretty sure Spotify will mess this up regardless of actual gender. I live in Germany but only speak English. Spotify can't work out what language to serve anything to me in.
Because the vast majority of their users are completely unaware and/or indifferent of this. You and other tech- and privacy-savvy people moving away from it will not hurt Spotify's bottom line at all.
I like that the campaign website complains about tracking and surveillance when Spotify does it but then uses Google Analytics on their own website fightforthefuture.org/.
The inventor listed on this patent left Spotify four years ago, and it clearly hasn't made it into any products released in the meantime. Not sure it's worth getting riled up about a patent filing — it could just be something the inventor wanted on his resumé.
Good for you, very individualist, looking out for yourself.
That said, the vast majority of people is not aware this is happening, will just hit 'agree' on anything that asks permission, etc. The only thing protecting them is legislation. It's up to people like you and me that are actually aware of these things to fight back.
I mean if it wasn't for 'us', tech- and privacy savvy people, secure and encrypted communications, privacy laws, etc would not be nearly as much of a thing that they are now.
That service you mentioned mentions that data will be exported to the US: https://www.di.fm/member/privacy. They don't seem to mention what user statistics or tracking they do though.
Well sure, but I don't believe signing a petition won't do anything. The only thing would be to stop paying them. I already have done that and I advice others to not be spotify customers as well.
They do say that under the "INFORMATION WE COLLECT AUTOMATICALLY" header. It is not much just the IP and some other basic stuff. It is a small company that runs the service so I don't think they have the resources that are required to do a lot of data collection and analysis and they wouldn't benefit from it either.
If Spotify holds a patent for this thing, does it mean that other companies wont be able to do the same thing? If yes, I'm just going to cancel my subscription and use another streaming service.
Some thoughts - each clause is independent of the others.
I am typically sympathetic to fighting over-aggressive surveillance for profit; I am wary of it more when govt uses it, but it can be dangerous with companies too. This is a valid concern.
The creators of the site clearly had a chip on their shoulder before this patent. It could read like they are "just mad" at Spotify for a number of reasons, and that this isn't a big deal. I think the EFF does better at selling the problems of seemingly benign tech.
I am beginning to get allergic to appeals to trans-this and racist-that when it seems so tangential: This isn't some systemic racist policy of government where the intent is muddy but the effect is clear - this is a creepy program that is made to make money on targeting services, whether you're trans or not.
>Emotion recognition software is largely seen as racist pseudoscience
It's like they're trying to derail the conversation they started. Why do so many petitions just throw a bunch of arguments in there and hope one of them sticks?
This petition is the sort of thing I'd come up with if I was asked to build a false flag operation to discredit people that care about privacy.
I'm no defender of surveillance, but this website is a highly paranoid take on what is just a patent application.
Sometimes companies file patents on dark-pattern type techniques just to prevent other companies from using those techniques.
Also, I guess the website creators don't realize this, a lot of speech recognition can be done completely on the device with the audio and the recognized words never leaving the device. (Modern mobile OS versions and increasingly desktop OS versions are giving the end user more and more control over whether they allow anything beyond this, on a per-app basis.)
The service could, for example, boil everything down to just one bit of information: The user likes this song. Or does not like this song. True, or false. One bit. That's pretty far from the drama presented on the website.
I still don't want my music player listening to my conversations or using my mic at all. And one bit can turn into 4 which can turn into streaming all my conversations to a cloud service to analyze them when management and developers turn over at a company.
This seems like a sensational ad for this guys new album, using fear-mongering to drive clicks and fame. It's a patent, not a feature. Besides, if they are so against surveillance capitalism, why use Google Analytics?
I use spotify ~8 hours a day during the work week. ~1750 hours of streaming last year (music, podcasts, etc.)
Two days ago, my iPhone updated the Spotify app. When I opened the app, the very first popup was "You can now use Spotify's voice controls! Say 'Hey Spotify' to try it out."
The documentation linked to the banner was all about how Spotify intends to start recording ambient audio from the phone's microphone, including what I assume would be any conversations.
I don't specifically agree with the sentiment of this website, but the idea of Spotify collecting audio is just another example of under-regulated tech companies willing to push ethical assumptions to their breaking point, all for the sake of making more revenue.
No. Spotify will begin receiving your voice input data when you press the action button or say the wake word or (for voice ads) when you hear an audible tone, and continue until Spotify has processed your question or request.
The device will always indicate to you when Spotify is receiving your voice input, for example with a visual indicator or an audible tone.
When listening for the wake-word, Spotify listens in short snippets of a few seconds which are deleted if the wake-word is not detected."
When we make material changes to this Policy, we'll provide you with prominent notice as appropriate under the circumstances, [...] We may notify you in advance.
Please, therefore, make sure you read any such notice carefully.
If you want to find out more about this Policy and how Spotify uses your personal data, please visit the Privacy Center on spotify.com to find out more.
113 comments
[ 3.1 ms ] story [ 165 ms ] threadTime to support and use alternative, free/libra, self-hosted solutions. https://funkwhale.audio is one off top of my head for Spotfiy.
Spotify (EDIT: for some people) solves the content problem for a low fee. Self hosted requires either spending lots of money, or in lots of regions something that probably qualifies as copyright infringement.
(EDIT: noted that this isn't universally solved, and thanks for responses pointing this out)
> Disk and USB space are cheap
Apologies for any ambiguity. When I was refering to spending lots of money, I wasn't considering the costs of self hosted infrastructure (servers, backups), but rather the cost of acquiring the media in the first place.
The solution is to firewall the internet in the UE.
I mean Spotify has been one of the biggest driving forces in reducing music piracy, in part thanks to being more convenient.
On the other hand, they don't pay artists nearly enough so the flipside is that to the artists it doesn't actually make that much of a difference whether someone listens to them on Spotify or just pirates the music.
Well, yeah, most of the time? Can you name a unicorn that doesn't collect and use data en masse?
I don't think the author knows what a patent is, because I also don't get the impression that the point of this is make the idea public domain instead of owned by one company.
If your browser supports searching text you'll find it too.
The Whois on the site says that it was registered on April 5th of 2021 for 1 year and the customer is based out of Ontario, Canada. Tucows was the registrant, and they seem to be a generic handler of web domains, based out of Toronto. https://tucows.com/
Other than that, I can't really find out any additional info about the site or who is running it.
I'm sure other HNers are more adept at searching for these kinds of things, and I'd love to know more about the paper-trail of it.
https://www.whois.com/whois/stopspotifysurveillance.org
I'm not singing up for what seems like a mix of legit concerns and ... borderline conspiracy theory / ideas that maybe really don't help the cause.
I'm inclined to sympathize with a lot of this but I'm not getting on board an unknown / the whole crazy train.
Similar to the Amazon "wage cage" patent that spawned so many memes and bad press.
This went so against my ethos I quit and found another job.
It scanned the filesystem to find local media files to present for playback. It used P2P to legally share locally cached parts of DRM:ed Spotify tracks to other users in order to lower Spotify's bandwidth costs.
Did you also ban Skype?
It was well known at the time. How do you know it was not sharing other files it found?
Making use of of peer to peer protocols to distribute content is something I wish more companies did.
P2P lowers the barrier to entry for new players in a domain, because it allows us to bring a product to the consumers with potentially drastically lower cost of operating, because we can provide service to our users without the massive bandwidth costs to our business.
This is good for the consumer too. So don’t be mad at companies that use P2P technology. Be in favor of it. And demand better terms from your ISP instead. So that the internet can continue to allow for innovation and competition, instead of further centralizing all of the power to only the biggest companies and slowing down innovation and hurting competition.
There is widespread suspicion that Facebook Inc already does this (they deny it, but I've heard and experienced too many anecdata to believe them).
Speaking as someone who can at times be a targeted-ads apologist, conversation-snooping software is an abhorrent invasion of privacy that I can't imagine anyone would be happy with.
Why spread an easily testable lie? People have checked, they don't do this.
BTW, I didn't claim that they do this, only that I and others suspect they do.
Personally I've seen cases were a topic was discussed aloud which was not searched/browsed by either party and received ads very shortly thereafter. Extremely specific stuff, like you randomly talk about life insurance and then get a life insurance ad.
Source: https://www.lifewire.com/how-facebook-messenger-apps-drains-...
> Facebook and Facebook Messenger apps for iOS and Android devices consume a lot of battery life. Besides complaints from people worldwide, authorities and analysts conducted tests. They affirmed that both are battery hogs even when the apps aren't in use.
I'm not sure how one would be expected to notice a "large upload of audio" since most people don't monitor their network activity, and slowdowns get blamed on a bad connection. I'd also expect smart FB engineers would be able to figure out how to minimize network activity (eg; sampled, low-fidelity audio).
EDIT: To be clear, I don't mean to say, "they use lots of battery therefore they must be spying", just that lack of battery drain isn't convincing evidence that they're not doing some sort of conversational spying.
Unless you think they are 'being sneaky' by developing zero day exploits to break permissions boundaries and prevent logging on both Android and iOS.
What's more likely is that: (a) we're all a lot more predictable than we'd like to admit, and (b) you underestimate the multitude of data sources available to Facebook and piped over to its ecosystem. Just because an ad is on Facebook, doesn't mean Facebook is the one providing all the targeting data.
Pretty much, yes. Perhaps only one of the two operating systems, and perhaps not all versions.
I'm not confident they've done that, but it'd be hard to prove they haven't, and frankly it seems like the kind of thing FB might do (I'm sure not everybody would agree).
It's certainly possible that FB is just wildly superior at understanding/predicting people than humans are. Their targeting is, after all, really good.
I see this claim frequently being made in such contexts, but it's never explained. Spotify playing music that matches your mood hurts trans people?
> Queer indie-punk musician and Fight for the Future director Evan Greer has released a new music video for the song “Surveillance Capitalism,” ahead of the release of the album Spotify is Surveillance.
That raised a red flag for me. Why is this person's gender identity in any way relevant to the issue at hand?
In the same way the text calls out emotion recognition software as "racist pseudoscience" (then linking to an article about facial recognition) - which doesn't have much at all to do with the patent.
Bought to you by the "Union of Musicians" - the same people who claim Spotify make "enormous profits" (I wouldn't call them enormous) and want 1 cent per play.
They want Spotify to pay them, but they also want to damage Spotify's reputation? Talk about shooting yourself in the foot
Further to that, check out the Union of Musicians' subcommittees [1] - there's a "Classical Music Accountability" subcommittee (classical music = white supremacy?)... and a Police Abolition committee, of course.
[1] https://www.unionofmusicians.org/subcommittees
The group seems to be sidetracked with every single cause, a subcommittee for every minority group... which, to me, takes away from the main purpose of a Union of Musicians - getting paid/treated properly (as a whole).
Which just reminds me of how this webpage gets sidetracked from the main issue. The main issue should be: listening in on conversations is creepy.
Normal people notice that it’s a blunt weapon more than serious advocacy for the marginalized.
Machine learning detection of people’s gender in the real world is anywhere from annoying to distressing to trans people because all the training data is cis people and so they get matched with their AGAB. Can you imagine how terrible it would be to already hate and be super self-conscious about your voice only to have Spotify start telling you via recommendations that you don’t pass well enough. Yikes.
Although the patent may be creepy...
The text of the patent refers to media recommendation (not advertising) - contrary to the sibling comment where you claim it will result in ads for Dollar Shave Club, the patent does not once use the term "advertising". Sure, it could be used for that, but I don't think that's the main intention at all.
How are you going to know that Spotify has recommended you a song based on your gender? Are we assigning genders to songs now? Straight white cis males can like Aqua's "Barbie Girl", too.
This is an absurd take. Do you see no difference between Spotify recommending songs to a trans woman based on what men with similar music tastes in her area and age group are listening to and "assigning gender to songs?"
Spotify also typically recommends you playlists rather than songs and my home page is extremely gendered. I get playlists like "You Go Girl", "Girls Like You", "Women of Rock", "Fierce Femmes." A trans guy getting these kinds of recommendations would be ... well odd.
Like recommendation engines aren't magic -- they look at what people "like you" listen to where "like you" has all sorts of random demographic data as well as your personal patterns/history.
At the moment, but I think the idea behind this patent is that it can factor in your mood (and other factors, such as gender) into the recommendation engine as well as what you already like.
Though I don't see gender being much of a signal. My GF and I have exactly the same taste in music (well, she doesn't like Classical that much, and I don't like Musicals, and she's more into Goth/Industrial/Metal than I am, but it's a 90% match). Anecdata, but that's all I have.
It's maybe not that bad of the grand scheme of things, but it's inarguably racist.
A trans woman constantly getting ads for dollar shave club, Nivia for men, and Weezer is hugely invalidating simply because some machine learning algorithm says she sounds like a man. Good god imagine having to perform a stereotypical version of your gender identity for your phone just to get the recommendations you want.
If you like electronic music I can recommend Digitally Imported (https://www.di.fm/).
The best part is that I don't need to find new music since it's radio. Also it's like half the price of a spotify sub.
That said, the vast majority of people is not aware this is happening, will just hit 'agree' on anything that asks permission, etc. The only thing protecting them is legislation. It's up to people like you and me that are actually aware of these things to fight back.
I mean if it wasn't for 'us', tech- and privacy savvy people, secure and encrypted communications, privacy laws, etc would not be nearly as much of a thing that they are now.
That service you mentioned mentions that data will be exported to the US: https://www.di.fm/member/privacy. They don't seem to mention what user statistics or tracking they do though.
https://www.stopspotifysurveillance.org is using Cloudflare SSL!
I am typically sympathetic to fighting over-aggressive surveillance for profit; I am wary of it more when govt uses it, but it can be dangerous with companies too. This is a valid concern.
The creators of the site clearly had a chip on their shoulder before this patent. It could read like they are "just mad" at Spotify for a number of reasons, and that this isn't a big deal. I think the EFF does better at selling the problems of seemingly benign tech.
I am beginning to get allergic to appeals to trans-this and racist-that when it seems so tangential: This isn't some systemic racist policy of government where the intent is muddy but the effect is clear - this is a creepy program that is made to make money on targeting services, whether you're trans or not.
>Emotion recognition software is largely seen as racist pseudoscience
It's like they're trying to derail the conversation they started. Why do so many petitions just throw a bunch of arguments in there and hope one of them sticks?
This petition is the sort of thing I'd come up with if I was asked to build a false flag operation to discredit people that care about privacy.
Have we already forgotten that PRISM proved that data collection is outsourced to companies? They're basically one and the same.
Sometimes companies file patents on dark-pattern type techniques just to prevent other companies from using those techniques.
Also, I guess the website creators don't realize this, a lot of speech recognition can be done completely on the device with the audio and the recognized words never leaving the device. (Modern mobile OS versions and increasingly desktop OS versions are giving the end user more and more control over whether they allow anything beyond this, on a per-app basis.)
The service could, for example, boil everything down to just one bit of information: The user likes this song. Or does not like this song. True, or false. One bit. That's pretty far from the drama presented on the website.
is it because it is a successfull european company?
I use spotify ~8 hours a day during the work week. ~1750 hours of streaming last year (music, podcasts, etc.)
Two days ago, my iPhone updated the Spotify app. When I opened the app, the very first popup was "You can now use Spotify's voice controls! Say 'Hey Spotify' to try it out."
The documentation linked to the banner was all about how Spotify intends to start recording ambient audio from the phone's microphone, including what I assume would be any conversations.
I don't specifically agree with the sentiment of this website, but the idea of Spotify collecting audio is just another example of under-regulated tech companies willing to push ethical assumptions to their breaking point, all for the sake of making more revenue.
"Is Spotify recording all of my conversations?
No. Spotify will begin receiving your voice input data when you press the action button or say the wake word or (for voice ads) when you hear an audible tone, and continue until Spotify has processed your question or request.
The device will always indicate to you when Spotify is receiving your voice input, for example with a visual indicator or an audible tone.
When listening for the wake-word, Spotify listens in short snippets of a few seconds which are deleted if the wake-word is not detected."
https://www.spotify.com/us/legal/privacy-policy/plain/#12-ch...
We may occasionally make changes to this Policy.
When we make material changes to this Policy, we'll provide you with prominent notice as appropriate under the circumstances, [...] We may notify you in advance.
Please, therefore, make sure you read any such notice carefully.
If you want to find out more about this Policy and how Spotify uses your personal data, please visit the Privacy Center on spotify.com to find out more.
Except it doesn't say that at all.