Today, we’re unveiling Anonym Private Audiences: a confidential computing solution allowing advertisers to securely build new audiences and boost campaign results.
Today, we’re unveiling Anonym Private Audiences: a confidential computing solution allowing advertisers to securely build new audiences and boost campaign results.
It has been shown repeatedly that “differential privacy” can be exploited to de-anonymize the users whose data has been aggregated.
If you read Mozilla’s description of their Private Audiences system you immediately ask, “what happens if an advertiser has an audience comprising a list of ‘known opposition party supporters’ and generates a new ‘audience’ based on that profile? Do they then get an expanded list of opposition party supporters to target?” Yes of course they do, because that’s entirely the purpose of this system.
Waving their hands and saying it uses ephemeral machine learning models and differential privacy does not solve the inherent societal problems with allowing targeted advertising.