Privacy implications of browsers’ (mis)implementations of Widevine EME (2023)
hal.science
hal.science
I don't know what exactly causes this, since it's intermittent (the same web site doesn't always do it) and happens even with various ad and tracking blockers in place.
browser.eme.ui.enabled = false
media.eme.enabled = false
What Signal is doing is trying to get the system to restrict the content from the rest of the system. Which might work as a transient hack but doesn't actually work to protect the user when the system is adversarial, because Microsoft (the adversary) has the DRM private keys. Even some hypothetical DRM system which is effective in oppressing the user wouldn't prevent Microsoft from purloining the user's data whenever they want because they're the ones who make the DRM.
This is similar to HTTPS certificate chain of trust. The root signing authority needs to be trusted, but once you break that trust there's no going back. It is a self-regulating system.
This is a bit different to encrypted video where it is actually being encrypted off device.
Plus, what is Netflix even going to do? Stop supporting streaming on Microsoft platforms and then lose a bunch of subscribers for no benefit to themselves?
Also, the issue is that somebody is going to copy a ~30 fps video using screenshots without audio taken at an interval of ~0.2 fps? Nobody is going to do it that way.
The promise of DRMed content at the moment is that (technically) no one is able to do so as there are no backdoors into it, for nobody.
What happens outside of that path, before it's decrypted or after it's displayed, is beyond what DRM is meant to control.
In order for the device to display the content on the screen, some part of the device has to have access to the plaintext. That part of the device is controlled by Microsoft -- it's their DRM system, they wrote that code and have the keys to update it -- so they inherently have access to it. Saying "that part is outside the scope of the DRM" is just defining your way around the fact that they can still do it.
Or to put it another way, if some court orders Microsoft to extract some DRM-protected content, what do you think happens?
So while Microsoft controls the platform, that doesn’t automatically mean they can trivially extract plaintext content from DRMed streams, especially not without compromising the integrity of systems they themselves may not fully own.
And even that is assuming there is no way to remove the watermark, which there always is because multiple copies of the same video can't each be uniquely identifying without revealing what's different about them.
Meanwhile the inconvenience to paying customers is real when their stuff doesn't work, and every customer who pirates your stuff because the paid offering doesn't work for them is an actual lost sale.
It really does seem like the DRM vendors are taking them for a ride.
Let's be clear here. That's a fine point in the generic sense, but in the Signal situation there are no private keys and it's not really DRM.
Suppose a third party app wants to make screen captures. Windows prevents it, because otherwise it could do the same thing to Netflix and capture the video. The thing preventing the app from bypassing that constraint is DRM.
Whereas suppose Microsoft wants to distribute an update to the video rendering code in Windows. It will have access to the data on the screen because it's the thing converting it into pixels, so Microsoft signs the new code with their private keys and distributes it to your PC and it gets access to what's on your screen. Which they could also do with code designed to exfiltrate it.
Also if Microsoft wanted to bypass it they could just ignore the function call, they wouldn't have to do any clever workarounds.
Simpler software could satisfy web users.
Could reduce potential for surveillance and annoying distractions. Easier to audit and control.
IRC for chatting, ICQ for instant messaging (which didn't work because my ISP at that time used a strange firewall / proxy setup and IRQ wasn't able to get through), newsgroups as a kind of discussion board, picture viewers for all kind of image formats (like wise video players), real player (Buffering...:D) for streaming....and most importantly web browser(s) when you want to grab information from all around the world (but dang having only a 33.6kbit modem, was really a test of patience sometimes).
Oh! I forgot! WinAmp which can whip the llama's...yeah you surely know what it does whip! ;)
But broadly yeah, same
(For the browser part of the DRM setup, I use Chrome/Chromium, the violate-me-all-the-ways browser. For all other browser purposes, I use both Firefox, the violate-me-fewer-ways browser, and Tor Browser, the draw-fire-of-state-actors-but-thwart-techbro-actors browser.)
To get the content from a different source in a more user-friendly format, right?
https://news.ycombinator.com/item?id=44294402
So they use Firefox 99% of the time and then if they encounter the rare thing that requires DRM they treat it like toxic waste that has to run in an isolated sandbox, which doesn't need to be the same browser they use for anything else.
The only other sensible option is to get out the reversing toolkit and break the DRM.
Privacy is a sliding scale and the idea that it's absolute privacy or nothing else isn't helpful IMO. If people want that, they should use Tor as nothing comes close.
>or a new opensource browser
Brave browser fulfills that role.
No uBlock is a deal breaker. Chromium is stuck with the neutered uBlock Lite thanks to Manifest V3.
> Brave browser fulfills that role
Sure, and it's also funded by VC money. How long until the vultures start swooping in to get a return on their investiment?
Why? If a browser is able to performantly and accurately block ads, why should the exact extention matter.
>and it's also funded by VC money
Which allowed them to properly invest into building out the browser and search engine.
>return on their investiment?
Controlling the home page / search of a web browser is extremely valuable.