Thanks for explaining the fingerprinting considerations, Eero!

One quick correction: the reduced User Agent string 
<https://developer.chrome.com/en/docs/privacy-sandbox/user-agent/> does not 
reveal OS/platform version as your recently added text seems to suggest. 
Other than that, I am satisfied that the API does not elevate 
fingerprinting risk.

On Tuesday, April 4, 2023 at 1:11:08 PM UTC-4 [email protected] wrote:

> Hi,
>
> On Thurday, March 30, 2023 at 3:43:06 PM UTC+3Kaustubha Govind wrote:
>
> Would you be able to expand your Privacy Considerations section to address 
> whether the API is likely to expose any new fingerprinting surfaces?
>
>
> I have added Fingerprinting 
> <https://github.com/riju/backgroundBlur/blob/main/explainer.md#fingerprinting>
>  
> subsection to our  Privacy Considerations section.
>
> Particularly; I am interested in understanding:
>
>    1. Whether there is additional information about the platform (e.g. OS 
>    version) that can be gleaned by querying whether the platform supports 
>    background blurring. Note that with the User Agent Reduction 
>    <https://developer.chrome.com/en/docs/privacy-sandbox/user-agent/> 
>    work, we are attempting to limit default access to some of this platform 
>    identifying information.
>
> In short, there isn't but see the linked Fingerprinting subsection for 
> detailed analysis. 
>
>
>    1. How stable do we expect the blur to be across hardware 
>    configurations? It would be unfortunate if this turned into a problem 
>    similar to Canvas Fingerprinting.
>
> In Canvas Fingerprinting, sites can draw to a canvas (the source) as they 
> like and fingerprint the result which may be a bit different on different 
> platforms and platform versions. I assume that you refer to this (lack of) 
> stable result here.
> This is really not an issue with background blur. In the case of the 
> getUserMedia tracks, the sites has access to tracks which may have 
> background blur in effect and may allow background blur to be 
> disabled/enabled and to the resulting frames. However, sites have no 
> control on the source (what's in the field-of-view of a camera) so it is 
> not possible to reprocess the same frames on different platforms and 
> platform versions and to compare the results.
>  
>
> Perhaps one mitigation here is that the capability is gated behind the 
> getUserMedia() permission; which limits drive-by fingerprinting?
>
>
> That is the main gate but see the linked Fingerprinting subsection for 
> detailed analysis. 
>
> BR,
>  Eero
>

-- 
You received this message because you are subscribed to the Google Groups 
"blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/a/chromium.org/d/msgid/blink-dev/c2b4ee97-ba9e-486f-b4b0-028e79c98918n%40chromium.org.

Reply via email to