That's the only part I'm interested in. I've read this article - or something similar - before and it doesn't surprise me that these big tech companies want more control. What I don't understand is how this affects linux desktop?
Is it going to be that online services or websites or webapps can choose to require attestation? Whether you use this OS or that OS? Or are linux developers forced to change their open source software?
It'll affect you the same way reCAPTCHA does. A website decides to use "Google Cloud Fraud", and locks you out if you don't complete the "challenges" it shows. Your OS doesn't matter at all.
I am going to assume that this also destroys millions of AI agents and bot scrapers this time which is why some “AI Engineers” were complaining about this recently.
Well, this is how Google will kill all the scrapers on its search data.
> Fraud Defense leverages a sophisticated and adaptable risk analysis engine to shield against automated software. It is specifically designed to orchestrate trust for the agentic web, neutralizing malicious scrapers while welcoming legitimate AI agents.
I'm sure it'll block a whole bunch of awful scrapers but if Google doesn't hate a bot, it'll be able to pass.
Sounds like an anti-competitive play to build an AI moat. They'll introduce a sham "verification program" and only allow bots operated by entities large enough to sue them for anti-competitive conduct.
Does mass scraping need google for content discovery? Surely most sites contain a site map or index that would effectively self enumerate once you know the domain, which is more often than not publicly disclosed?
the only anticompetitive element I can think of is the way they pushed their scanning app to Android phones with Play Services. On IOS they're not in control but still able to launch an app (app snippets the feature is called, I think?) but on Android they themselves killed off Instant Apps because nobody used it. If one of Google's competitors like hCAPTCHA tries to do the same, they'll have more friction on Android than Google does.
When it comes to GrapheneOS, it's the website owners that decided to block those devices by using this service. There are other services that don't block those phones they can use instead.
That's the whole goal of the concept. Safetynet (the predecessor of Play Integrity) was developed to block CyanogenMod and then later used to block Huawei.
Sure, Google was betting that bureaucratic companies would enroll voluntarily and it worked.
> Like with reCAPTCHA, there are other services and libraries out there to detect root access and other things companies want to detect in their apps.
My opinion on this is that any method to check integrity, root access or if developer mode is enabled is a security vulnerability by itself, no such app should be able to know that.
> My opinion on this is that any method to check integrity, root access or if developer mode is enabled is a security vulnerability by itself, no such app should be able to know that.
I think knowledge of such information should be available to all apps, but I think apps should not be so annoyingly restrictive. There's absolutely no reason why games or generic apps need to act on any of this information.
Blind people need to start suing at least in America the ADA is far easier to win against large companies than in the UK as the equality act is treated vastly weaker by judges than if someone presented a religion or race lawsuit.
America is the only place to take down big tech discrimination.
Apple+Google got punished by the EU for non-competitive practices and now they offered to ordinary websites their most desired features: bot blocking and unavoidable user tracking across all devices and operating systems.
And if EU wants to sue, they'll have to sue each and every website that requires this, and they would loose, because there are no alternatives and even if there were, they would be just as bad.
If Windows wasn't so far behind Apple and the rest of the industry in regards to integrity APIs this wouldn't be necessary. It's embarrassing for Microsoft that someone needs to use a separate, more secure device since their security is so bad.
Attestation isn't against being able to do whatever you want with your own device. It just means that if you want other people to trust your custom device you need to get them to trust your signing key.
There are many changes that are possible which do not harm the integrity of applications.
>the user can’t do what they want with their own device
In the same way the user can't
make their device have the Microsoft Word app send them $1 million from Microsoft's bank account. Once other people are in the picture you can't always have your way.
Windows Hello offers an attestation API according to the releases I found, though because Microsoft has called at least four products "hello" now, I can't easily find the details. I don't think there's a technical reason why Google couldn't have released an app with a URL handler that uses that API except maybe for the Windows TPMs being less secure than mobile ones in general.
It seems like the documentation for the feature is aimed entirely at MDM setups, though.
The basic API requirements are all there, and Windows 11 requires TPM 2.0, so I believe it should be possible for Google to build a Play Integrity equivalent around that.
"strong integrity" also takes into account if a security update has been installed recently enough. I don't believe hardware integrity spoofing has been accomplished on Android yet. Software integrity and compatibility with old hardware has been used to spoof device IDs and pretend a phone doesn't have the ability to do hardware attestation.
It's technically possible to exploit a kernel and get root access on a running device, of course, but the persistent root that is used most often will be detected by hardware integrity mechanisms. Exploit based root might be as well if it makes itself detectable enough.
> if a security update has been installed recently enough
In turn, this enables any tyrannical or anti-competitive demand which can be implemented in software, such as "user is not on the blasphemer list" or "all communications are being CC'ed to the Ministry of Truth."
> "strong integrity" also takes into account if a security update has been installed recently enough.
My Galaxy S10, last update in 2023 passes strong integrity.
With the little amount of security updates most Android devices have, I'm pretty sure you can find an exploit for pretty much everything except the most expensive flagships.
What does integrity really means when nobody really knows what's in the device and with a terrible software update policy anyways.
The exact requirements for security updates depends on the Android version you're running and the one your device came with. From the docs:
MEETS_STRONG_INTEGRITY
The app is running on a genuine and certified Android device with a recent security update.
On Android 13 and higher, the MEETS_STRONG_INTEGRITY verdict requires MEETS_DEVICE_INTEGRITY and security updates in the last year for all partitions of the device, including an Android OS partition patch and a vendor partition patch.
On Android 12 and lower, the MEETS_STRONG_INTEGRITY verdict only requires hardware-backed proof of boot integrity and does not require the device to have a recent security update. Therefore, when using the MEETS_STRONG_INTEGRITY, it is recommended to also take into account the Android SDK version in the deviceAttributes field.
A single device will return multiple device labels in the device integrity verdict if each of the label's criteria is met.
The S10 should be on Android 13, so it should not pass STRONG_INTEGRITY. If it does, perhaps it's possible Google updated the docs early in anticipation of a change? The software update requirement wasn't always there.
> linux desktop
That's the only part I'm interested in. I've read this article - or something similar - before and it doesn't surprise me that these big tech companies want more control. What I don't understand is how this affects linux desktop?
Is it going to be that online services or websites or webapps can choose to require attestation? Whether you use this OS or that OS? Or are linux developers forced to change their open source software?
It'll affect you the same way reCAPTCHA does. A website decides to use "Google Cloud Fraud", and locks you out if you don't complete the "challenges" it shows. Your OS doesn't matter at all.
Thanks! It was slightly confusing, now I get it.
I am going to assume that this also destroys millions of AI agents and bot scrapers this time which is why some “AI Engineers” were complaining about this recently.
Well, this is how Google will kill all the scrapers on its search data.
Not entirely, Google's own page says:
> Fraud Defense leverages a sophisticated and adaptable risk analysis engine to shield against automated software. It is specifically designed to orchestrate trust for the agentic web, neutralizing malicious scrapers while welcoming legitimate AI agents.
I'm sure it'll block a whole bunch of awful scrapers but if Google doesn't hate a bot, it'll be able to pass.
Sounds like an anti-competitive play to build an AI moat. They'll introduce a sham "verification program" and only allow bots operated by entities large enough to sue them for anti-competitive conduct.
Does mass scraping need google for content discovery? Surely most sites contain a site map or index that would effectively self enumerate once you know the domain, which is more often than not publicly disclosed?
Aside from the horrendous privacy implications, is there a possible argument that this is anti-competitive?
the only anticompetitive element I can think of is the way they pushed their scanning app to Android phones with Play Services. On IOS they're not in control but still able to launch an app (app snippets the feature is called, I think?) but on Android they themselves killed off Instant Apps because nobody used it. If one of Google's competitors like hCAPTCHA tries to do the same, they'll have more friction on Android than Google does.
When it comes to GrapheneOS, it's the website owners that decided to block those devices by using this service. There are other services that don't block those phones they can use instead.
That's the whole goal of the concept. Safetynet (the predecessor of Play Integrity) was developed to block CyanogenMod and then later used to block Huawei.
App developers need to put effort into enabling these APIs so it's not like Google is actively blocking your favorite apps. Their makers are.
Like with reCAPTCHA, there are other services and libraries out there to detect root access and other things companies want to detect in their apps.
Sure, Google was betting that bureaucratic companies would enroll voluntarily and it worked.
> Like with reCAPTCHA, there are other services and libraries out there to detect root access and other things companies want to detect in their apps.
My opinion on this is that any method to check integrity, root access or if developer mode is enabled is a security vulnerability by itself, no such app should be able to know that.
> My opinion on this is that any method to check integrity, root access or if developer mode is enabled is a security vulnerability by itself, no such app should be able to know that.
I think knowledge of such information should be available to all apps, but I think apps should not be so annoyingly restrictive. There's absolutely no reason why games or generic apps need to act on any of this information.
So fuck blind people I guess?
Blind people need to start suing at least in America the ADA is far easier to win against large companies than in the UK as the equality act is treated vastly weaker by judges than if someone presented a religion or race lawsuit.
America is the only place to take down big tech discrimination.
That is a cost that our future authoritarian world leader has decided is more than acceptable.
So, let me see if I understand it:
Apple+Google got punished by the EU for non-competitive practices and now they offered to ordinary websites their most desired features: bot blocking and unavoidable user tracking across all devices and operating systems.
And if EU wants to sue, they'll have to sue each and every website that requires this, and they would loose, because there are no alternatives and even if there were, they would be just as bad.
Great job Google+Apple! I'm proud of you. /s
If Windows wasn't so far behind Apple and the rest of the industry in regards to integrity APIs this wouldn't be necessary. It's embarrassing for Microsoft that someone needs to use a separate, more secure device since their security is so bad.
It's embarrassing for Hacker News that people here are commenting to support attestation systems that prevent you from owning the device you bought.
Attestation isn't against being able to do whatever you want with your own device. It just means that if you want other people to trust your custom device you need to get them to trust your signing key.
Pray tell, how might you get them to trust your signing key? Do you just email Mr. Pichai and ask nicely, is that enough?
Not sure if you’re being deliberately obtuse, but a signing key means nothing by itself. What exactly do you think is being attested TO?
Thats right: that the user can’t do what they want with their own device. Obviously your key wouldn’t be trusted if they could.
There is no other conceivable purpose that attestation could serve.
> Not sure if you’re being deliberately obtuse
Yes, they are. If there's a thread on HN about user-hostile features, you can be pretty confident that they've written a comment defending it.
There are many changes that are possible which do not harm the integrity of applications.
>the user can’t do what they want with their own device
In the same way the user can't make their device have the Microsoft Word app send them $1 million from Microsoft's bank account. Once other people are in the picture you can't always have your way.
Windows Hello offers an attestation API according to the releases I found, though because Microsoft has called at least four products "hello" now, I can't easily find the details. I don't think there's a technical reason why Google couldn't have released an app with a URL handler that uses that API except maybe for the Windows TPMs being less secure than mobile ones in general.
That attestation is for attesting you are using a TPM for user authentication. Which is different than attestation of integrity.
They do have some kind of attestation mechanism to actually attest the device state: https://learn.microsoft.com/en-us/azure/attestation/tpm-atte...
It seems like the documentation for the feature is aimed entirely at MDM setups, though.
The basic API requirements are all there, and Windows 11 requires TPM 2.0, so I believe it should be possible for Google to build a Play Integrity equivalent around that.
Integrity doesn't guarantee any security to your device, just that the device is same as from the factory. That's a common misconception.
"strong integrity" also takes into account if a security update has been installed recently enough. I don't believe hardware integrity spoofing has been accomplished on Android yet. Software integrity and compatibility with old hardware has been used to spoof device IDs and pretend a phone doesn't have the ability to do hardware attestation.
It's technically possible to exploit a kernel and get root access on a running device, of course, but the persistent root that is used most often will be detected by hardware integrity mechanisms. Exploit based root might be as well if it makes itself detectable enough.
> if a security update has been installed recently enough
In turn, this enables any tyrannical or anti-competitive demand which can be implemented in software, such as "user is not on the blasphemer list" or "all communications are being CC'ed to the Ministry of Truth."
>I don't believe hardware integrity spoofing has been accomplished on Android yet.
It has, but extracted keys aren't free.
> "strong integrity" also takes into account if a security update has been installed recently enough.
My Galaxy S10, last update in 2023 passes strong integrity.
With the little amount of security updates most Android devices have, I'm pretty sure you can find an exploit for pretty much everything except the most expensive flagships.
What does integrity really means when nobody really knows what's in the device and with a terrible software update policy anyways.
The exact requirements for security updates depends on the Android version you're running and the one your device came with. From the docs:
The S10 should be on Android 13, so it should not pass STRONG_INTEGRITY. If it does, perhaps it's possible Google updated the docs early in anticipation of a change? The software update requirement wasn't always there.