Event-level reports should indicate how noise has been applied

[csharrison]

We should consider adding fields like:

metadata_noise_probability or randomized_response_probability

This will make it easier to handle multiple implementations with different probabilities (or different experiments in the same implementation). It also signals to developers that values may be noisy.

cc @maudnals @johnivdel

[apasel422]

I'm happy to spec and implement this.

[apasel422]

Do we want to have separate fields for source event ID noise probability and trigger data noise probability?

[csharrison]

Do we want to have separate fields for source event ID noise probability and trigger data noise probability?

Yes I think that should be fine for now.

[apasel422]

@csharrison Are we concerned about the potential for these values to be used for fingerprinting?

[csharrison]

@csharrison Are we concerned about the potential for these values to be used for fingerprinting?

No I am not concerned. Users of the API need to know what the API is doing, so I think defending against this is going to be very challenging and we won't be exposing anything more than browser major version anyways (in majority of cases) which is being sent by default:
https://developer.chrome.com/docs/privacy-sandbox/user-agent/