Meta Uses VRS AI System To Combat Prejudice Based On Ads

When housing discrimination was permitted on its platform through tailored adverts, Meta was summoned to court back in 2019.Using the "Special Ad Audiences" tool, the tech giant was accused of enabling advertisers to break the Fair Housing Act.

When Meta replaced the aforementioned feature with a new AI system named "Variance Reduction System," the dust from that litigation was ultimately cleared up last year (VRS).

Now that VRS is in place in the US, let's fast forward to the present. Meta has revealed this.

This innovative algorithm modifies the advertising system in a way that enables advertisers to connect with their intended market without unintentionally discriminating against other demographics.

VRS is Meta's answer to ad-based prejudice.

Once an advertisement has reached a predetermined threshold of users, the AI system kicks in and begins to work.

It now compares the desired demographic of the advertiser with the demographic makeup of all users reached.

In cases where there is a mismatch or space for improvement, VRS will raise the ad auction price in order to reach a larger audience and finally reach the intended demographic.

Overall, this new ad algorithm compels the advertiser to change their bids in order to reach a broader pool and, consequently, reach their target audience.

Another way to describe it is that companies may no longer exclude specific social groupings from their advertising operations.

When we discuss how VRS analyses user demographics, the issue of privacy inevitably comes up.

The "noise" technology that Meta has incorporated into this algorithm to combat this issue keeps individual information from falling into the AI's grasp.

VRS currently only functions with housing advertising that are shown in the US. Ads for credit and jobs are also expected to be included in the new system.