hckrnws

> "Apps were automatically taking screenshots of themselves and sending them to third parties. In one case, the app took video of the screen activity and sent that information to a third party.”

> Out of over 17,000 Android apps examined, more than 9,000 had potential permissions to take screenshots. And a number of apps were found to actively be doing so, taking screenshots and sending them to third-party sources.

Which permission is that, and how do you detect which apps are doing that and stop them?

There is a permission to record the screen. It requires user consent and there's an icon in the status bar while it's being used. It's impossible to use this covertly.

What I believe the article is speaking about, is an app taking screenshots of its own windows. This is obviously possible and obviously requires no permissions whatsoever. Just make a screen-sized bitmap and do

    getWindow().getDecorView().draw(new Canvas(bitmap));

> It's impossible to use this covertly.

*Unless there's a zero-day that allows it.

If you're going to exploit a privilege escalation vulnerability from your app, why not just grab the most interesting parts of the /data partition while you're at it?

Burning a zero-day like that for targeted advertising seems extremely unlikely to me.

I followed the links to the study they referenced, and it says:

> Unlike the camera and audio APIs, the APIs for taking screenshots and recording video of the screen are not protected by any permission

However they also talk about doing static analysis on 9,100 out of the 17,260 apps, to determine (amongst other things) “whether media APIs are actually referenced in the app’s code”.

They then talk about doing a dynamic analysis to see which apps actually call the APIs (rather than just link to a library that might call it, but the app never calls that function the library).

The soundbite is bad, it shouldn’t say “had potential permissions to take screenshots”, it should just say “had the potential to take screenshots”

I doubt there's a specific "ability to send surreptitious screen shots to developer" permission. It must be a combination of permissions: one for making network connections, another for capturing the screen without making it obvious to the user, etc.

When it's a developer tool we call it RUM or real user monitoring. It's super useful for solving bugs, but obviously the potential for abuse or user hostile activity is super high.

... and is this permission to take screenshots of anything else you are doing on your phone at any time, or is it permission to take screenshots while you have that app open?

BTW, "smart" TVs send screenshots too. [0]

[0] https://dl.acm.org/doi/10.1145/3646547.3689013

We’ve reached the state where you can safely presume anything “smart” is violating your privacy.

Anything network connected.

The thing is, it's not even people doing the correlations. Just like transformers can learn most of human knowledge just by trying to predict tokens, I would not be surprised if the ad-serving machine learning systems have learned about people in similar detail.

State of the art about 10 years ago was 4 9s of accuracy predicting click-through rates from the available context (features for user profile, current website, keywords, etc.), which I interpreted as requiring a fairly accurate learned model of human behavior. I got out of that industry so I don't know what current SOTA is for adtech, but I can only imagine it is better. The models were trained on automatically labelled data (GB/s of it) based on actual recent click-through rates so the amount of training data was roughly comparable to small LLMs.

Recent anecdote; three of us were sitting around the kitchen table with our phones out chatting about an obscure new thing that had come up; it appeared in one of our FB ad streams pretty quickly.

My top guesses about how this is possible today;

1) Apps routinely link many third-party data gathering and advertising libraries. Any of these libraries could be gathering enough contextual data and reselling it to make a correlation possible. It's not just obscure thing A that triggers an ad, it's highly correlated mixtures of normal things X, Y and Z that can imply A.

2) other friends may have talked about the obscure thing recently and social network links implied we would be aware of it through them.

Distant 3) the models are actually good enough to infer speech from weird side-channels like the accelerometer when people wave their hands when they talk, etc. Accelerometer sample rate is < 1KHz but over 100Hz which may be enough, especially when you throw giant models at it.

> an obscure new thing that had come up

Since you've provided no explicit counter-evidence, I'm gonna go ahead and say I have four nines of accuracy in predicting that your smartphone was squarely in the dependency chain of any "obscure new thing" you could have imagined discussing.

Edit: wording

At one of my previous companies we made a moderately popular mobile app SDK that app developers would embed in their apps. We were approached by a company that claimed they had a MIT developed (or was it Bell Labs?) audio recognition technology similar to Shazam, but orders of magnitude more efficient, that would be used to recognize audio from ads and record when a user was exposed to a TV or radio ad for tracking purposes.

I don’t remember the name, that was at least 10 years ago before Apple started enforcing permissions on microphone access and showing an orange dot, but they wanted to do a revenue-share deal in exchange for us quietly bundling their SDK inside ours.

Needless to say we turned them down so we never learned more or tested the veracity of their claims, but there are some really sleazy companies out there. Modern smartphones have sufficient horsepower to do the audio processing on-device so the argument that this would show up in network traffic does not hold.

> There is no easy way to close this privacy opening

Sure there is.

Hide screenshot taking behind permission and slap down hard apps that refuse to operate without them.

It says "screenshots of themselves". The application is responsible for rendering the screen in the first place so it fundamentally doesn't need a permission.

Now, what could reasonably be a permission is "access the internet", but our overlords don't approve of that thought.

(Contrast this to web pages, which do not render themselves and thus can sensibly be blocked from screenshotting)

I mean yeah technically the website can’t screenshot, but it can do many functionally equivalent things.

For example, it can capture the entire DOM and send it off, including the contents of input fields that have not been submitted.

That DOM capture can be replayed on a browser to show what the user sees. So what’s the difference?

Well, blocking javascript would stop that. Noscript is a thing that some people use.

For an increasing plurality (possibly even majority at this point) of sites where the purpose is not purely to read text, this is effectively equivalent to saying "you can just not use the site."

All I/O (including timing, date/time, internet, and everything else) should be behind permissions (although some may be permitted by default, they should still be overridable). Furthermore, all I/O should allow the user to program proxy capabilities (which can be used for testing error conditions, as well as for privacy and security, and for finer permissions, and logging, and other stuff).

However, if an app wants to make a screenshot of itself, then it could do so by emulation of itself (so no permission is needed), as long as everything it displays is rendered by its own code rather than calling other functions in the system to do so.

> As far as anyone could understand, the proposed CMG system wasn't listening through a phone's microphone 24/7, instead it was using those small slivers of voice data that are recorded and uploaded to the cloud in the moments after you activate your voice assistant with a "Hey Google" or "Hey Siri" command.

That's not quite accurate. The CMG thing was very clearly a case of advertising sales people getting over-excited and thinking they could sell vaporware to customers who had bought into the common "your phone listens to you and serves you ads" conspiracy theory. They cut that out the moment it started attracting attention from outside of their potential marks. Here's a rant about that I originally posted as a series of comments elsewhere: https://simonwillison.net/2024/Sep/2/facebook-cmg/

The "Hey Google" / "Hey Siri" thing is a slightly different story. Apple settled a case out of court for $95m where the accusation was that snippets of text around the "Hey Siri" wake word had been recorded on their servers and may have been listened to by employees (or contractors) who were debugging and improving Siri's performance: https://arstechnica.com/tech-policy/2025/01/apple-agrees-to-...

The problem with that lawsuit is that the original argument included anecdotal notes about "eerily accurate targeted ads that appeared after they had just been talking about specific items". By settling, Apple gave even more fuel to those conspiracy theories.

I wrote about this a few months ago: https://simonwillison.net/2025/Jan/2/they-spy-on-you-but-not... - including a note about that general conspiracy theory and how "Convincing people of this is basically impossible. It doesn’t matter how good your argument is, if someone has ever seen an ad that relates to their previous voice conversation they are likely convinced and there’s nothing you can do to talk them out of it."

... all of that said, I 100% agree with the general message of this article - the "truth is more disturbing" bit. Facebook can target you ads spookily well because they have a vast amount of data about you collected by correlating your activity across multiple sources. If they have your email address or phone number they can use that to match up your behaviour from all sorts of other sources. THAT's the creepy thing that people need to understand is happening.

"Convincing people of this is basically impossible. It doesn’t matter how good your argument is, if someone has ever seen an ad that relates to their previous voice conversation they are likely convinced and there’s nothing you can do to talk them out of it."

It sounds more like we have evidence of what we believe, you think we should toss the evidence for your counter-theory, and people won't do that. We also have an effect where tons of people experienced this. You want us to toss that, too.

"You don’t notice the hundreds of times a day you say something and don’t see a relevant advert a short time later. You see thousands of ads a day, can you remember what any of them are?"

On Facebook, during one period this happened, they were only showing me adds for Hotworx and a massage place every time. Trying to stay pure minded following Jesus Christ means I avoid such ads. So, it was strange that it's all they showed me. Then, strange the only break from the pattern was showing unlikely topics we just talked about in person.

So, I'm going to stick with the theory that they were listening since it best fit the evidence. I don't know why they'd do it. Prior reports long ago said they used to use ML (computer vision) to profile people outside of the platform who showed up in your pics.

I'll note another explanation. Instead of always listening, they could have done it to a random segment of people who were rarely clicking ads. Just occasionally, too. We wouldn't see the capability in use all the time. A feature tested or used on a subset of users.

Also, these companies keep saying on us in increasingly creative and dishonest ways. If anyone is to be blamed, it's them.

Thank you for illustrating my point so perfectly.

My younger bro is convinced phones are eavesdropping on conversations and got particularly paranoid (I thought) a year or so back in regard to talking in earshot of his phone.

His evidence is empirical - Apparently he gets pretty high with friends and shit talks - but when when the search started to suggest some pretty way out things along the same lines, he landed that their conversations weren't private any more.

So I have an understanding of how much tracking is going on so I pressed him on that. But he assured me it was stuff he would not even bother to look up in a clearer mindset and of course smoking recreationally for a very long time knows not to go near some tools that could land himself trouble or awkward explanations. That's probably true he says a lot of stuff that a half decent search would put him straight. In the end I just figured loose permissions of one of the many apps he's installed and that's how they (the app) make their money, selling illegally obtained data to more legal sources.

Permissions are the problem with android phones - there needs to be a specific install route for users, one that the app starts asking for things it should not need have access to, the installer refuses to install and suggests the user look for something better. Camera apps for example really don't need access to communication channels, if it's updates it's need, it can ask - one time access.

Something I discovered when going down this rabbit hole is that if you had that conversation in your house and your visitors have access to your wifi, it may be that they performed the search without you knowing, and your ISP connected that data to you and sold it (as they do).

Location location location.

- User 1 shows an interest in <topic>.

- User 1 visits the same location, for the same period of time, as user 2.

- So I show an ad for <topic> to user 2.

How would your ISP connect that data if every search engine uses HTTPS now, so there's no way for the ISP to see what you were searching for?

DNS lookups are still frequently in the clear, and even if they're not, that just means you're trusting some DNS-over-HTTPS provider. The incentives are perverse.

And of course whoever you are performing your search with, like, oh, an ad company like Google, Meta, or Facebook? They just might use that search data for something.

Exactly. Google or Meta can correlate behavioral data like this. Your ISP cannot do that by intercepting your searches.

I care about accuracy when it comes to privacy conversations. I don't want people wasting their time on theories that aren't true when they should be focusing on the real issues at stake.

Yeah, it's Google and Facebook - not the ISP.

That's true. I had to rule that out by only counting instances when my friends and I were alone. If not, or Wifi is open, then who knows.

> Apparently he gets pretty high with friends and shit talks - but when when the search started to suggest some pretty way out things along the same lines, he landed that their conversations weren't private any more.

I had an experience like this several years ago. I was having dinner with a customer, and one of the guys brought up this story about how he went to school with someone who got caught cheating on Who Wants to be a Millionaire. Later, back at my hotel, I pulled up YouTube and the first recommended video was of the guy who got caught cheating on the game show. I had not searched for this during the conversation (or prior) nor do I watch game show videos on YouTube, or cheating scandal videos on YouTube.

Here's what I think happened: somebody at the dinner googled it, and the video got recommended based either on geo-location data (we were in close proximity) or because the person who googled it was in my phone contacts, or maybe both. But, I don't think Google/Youtube was recording anyone's conversation to make that recommendation.

It could also be that YouTube started recommending this video to people for whatever reason, which was why it was on this guy’s mind.

Anything is possible, but he didn't start the conversation about cheating. Someone else brought up something to the effect of they thought game shows were fake, then he told his story and a third person the table searched for and showed the video.

> Permissions are the problem with android phones - there needs to be a specific install route for users, one that the app starts asking for things it should not need have access to, the installer refuses to install and suggests the user look for something better. Camera apps for example really don't need access to communication channels, if it's updates it's need, it can ask - one time access.

I definitely don't want my phone making those decisions for me; I want my phone enabling me to make decisions. The app asks for permissions, I say no, and, rather than ratting me out to the app, my phone does its best to pretend to the app that it (the app) has the permission it wants, say by giving an empty contact book or whatever. (I know rooted phones can do this, but it shouldn't have to be something I have to fight my phone for.)

He is right, all modern phone brands are surveillance devices furnished to provide the OEM with identifying data: https://arstechnica.com/tech-policy/2023/12/apple-admits-to-...

He’s right and everyone knows it. It's pretty blatant and there have been lawsuits settle rather than go to a trial that would surely reveal the extent to which this thing that’s obviously happening is happening

https://www.sfchronicle.com/bayarea/article/apple-siri-priva...

I attempted to debunk that one here (an admittedly impossible task but I can't help myself trying): https://simonwillison.net/2025/Jan/2/they-spy-on-you-but-not...

A swan can't stop a hurricane

OK wow that actually fits here. https://simonwillison.net/2025/Apr/23/meaning-slop/

It is irrelevant. The suggestion that spying is for advertisement makes no difference.

That idea only exists to create fake two-dimensional anti-capilist rethoric, which is a rethoric easier to put down than the fact that privacy does not exist anymore.

So, I am supposed to do this. To "correct you" and look very lunatic.

It serves, however, a very specific goal. First, it cannot be copied en masse. If this behavior is copied (even as a meme), it implies doom to the more easier to defeat anti-capitalist rethoric and the birth of a true 3D anti-capitalist rethoric. It can only be mocked (smoking guy pointing to a conspiracy board), but that mockery is getting real serious real fast now.

Can I dive deeper into the mechanics of how this is gonna go?

We had so many chances, of doing good. You all had so many chances.

There's a nation proud of overspinning enrichment turbines with a complicated computer virus that can even work offline. No conspiracy, that's just StuxNet.

So, when you start learning about tech, you get paranoid. If you're not, it's even weirder.

The fact that someone can target you, individually, is undisputable. Whether it will or not, that's another question.

What I can recommend if you think you are being observed, is to avoid the common pitfalls:

Don't go full isolationist living without technology. That is a trap. There is nowhere to hide anyway.

Strange new friends who are super into what you do? Trap.

You were never good with girls but one is seemingly into you, despite you being an ugly ass dirty computer nerd? That is a trap. Specially online but not limited to it.

Go ahead, be paranoid. When an article comes to probe how paranoid you are, go ahead and explain exactly how paranoid you have become.

But live a normal life nonetheless, unaffected by those things. Allow yourself to laugh, and be cool with it.

Hundreds of clone accounts doxxing me? Well, thanks for the free decoys.

Constant surveillance? Well, thank you for uploading my soul free of charge to super protected servers.

Dodgy counter arguments in everything in care to discuss? Sounds like training.

The paranoid optimist is quite an underrated character. I don't see many of those around.

Sounds like the age old adage: if it's too good to be true, it is.

I also tend to be very skeptical towards popular sayings. Sometimes, they fail.

"true" in the sense you used here. Have you thought about what it means in that context?

We live in an age full of fear of missing out baits and reversed versions of such. There is no sense of "oh, this is good for me" that can be relied upon (implied in the original comment, you are going to find it), although there are sayings.

If it sounds too good to be true, it probably is. Otherwise it's just a tautology.

[flagged]

There is a list of things I keep under profound consideration always.

Information that travels backwards in time is one of them. I have a pretty good idea on how it could be possible and who would have the resources to do it.

God is also another. However, I am a very unorthodox student of religion. I deeply respect anyone that uses it to foster a good behavior. Whoever uses it to trick others, I tend to see more as an act of hostility towards innocent believers. Like, if someone tries to put me into a religion mindset just to fuck with me, it's a dick move.

What I know for sure is that God would not make mistakes. Whatever monitors me, does. It did so many times. I know it embarrasses them. It's delightful in that sense. So, yeah. God might exist, but I ain't talking about it when I describe paranoia.

Another thing that is quite recent in my studies is psychology and how we are all so vulnerable to it. I started to despise it a little bit. How come it never solved so many issues? How come it seems to put them to evidence but not fix them, and by putting them to evidence, make them worse?

Anyway. Do you want even more paranoia? If you like it, I should be supposed to charge for it, you know.

Doesn't it have to listen to everything to capture the wake word "hey siri"? How else is it done?

The iPhone has dedicated low-power on-device hardware that is trained to pick up "Hey Siri" exclusively. It only wakes up the rest of the device and captures additional audio after that wake word has been triggered.

https://machinelearning.apple.com/research/voice-trigger

https://machinelearning.apple.com/research/hey-siri

I seem to recall that state of the art audio encoding can compress voice to 8kbit/s which is a single packet per second, insignificant compared to how chatty your device is. Trivial to buffer and send during a period of activity. It sums to 1.7MB over the 30 minute window in the article graphs which should be visible if it is actually counted. Why would apple or google actually make it count though? They want to spy on you either for their own benefit or because the government forces them to. You say you found it taking screenshots and phoning them home. Of course! It is a surveillance device. Is it worse? Maybe. You should consider it sends everything home. Every keystroke, every touch of the screen, every sample of the accelerometers, every sample of audio. Perhaps only the sheer quantity of data in video prevents them from sending it all. Might be "remedied" with 5G bandwidth.

Audio, screenshots, and some of the other stuff I can believe, but I think batteries need a big upgrade before the data snatchers can get away with streaming video, even at a low bitrate.

I'm also not sure how easy keylogging is these days, is there even a permission that allows it? I supposed there's ways to do it with custom keyboards. Google/Apple doing it themselves would be a pretty big deal.

I think everyone acknowledges that chrome sends every keystroke in the address bar home. I don't keep up with the spyware so perhaps it is now every keystroke in the rest of the browser. It isn't much of a leap further that their operating system does the same.

Knowing how digital advertising works, it's more likely that a payload is delivered to the phone in some app or by os or by browser that has a dictionary of keywords paid for to be associated with specific ad campaigns. If the device detects that term (via sound, search, or media) it triggers a message home as an analytics to target you and your device now calls for those campaigns.

If it works like that, why aren't the app companies describing exactly how it works to advertisers in order to earn their business?

They describe how everything else they do works in great detail if you're someone who buys ads.

What makes you think the raw audio stream needs to be sent anywhere. Modern phones are capable of doing keyword extraction on-device.

You need to know what keywords to listen for before discarding the audio data. An advertising giant might know but a government doesn't.

This conspiracy theory has been around for a lot longer than phone hardware has been capable of doing that.

The Chrome Browser can transcribe audio into text, with what I consider good accuracy. It's well out of the realm of a conspiracy theory when it's been demonstrable for a couple decades.

Don't forget energy usage. The phone would need to be on high power mode all the time to run those kinds of algorithms. There's a reason "Hey Siri" has dedicated low-power hardware - it means it can work without burning through the battery.

If that were true why are cell phone voice calls still so terrible?