Facebook Deploys Machine Learning To Sniff Out Discriminatory Advertising

In no uncertain terms, Facebook says "discriminatory advertising has no place" on the world's largest social network. To keep it off of there, Facebook last fall started providing better education to advertisers about its prohibition against discrimination. It also beefed up the procedures it uses to enforce prohibition. Now it is taking things a step further by implementing strong enforcement tools and testing machine learning solutions.

This is not a new frontier for Facebook, or technology companies in general. Machine learning is one of the hottest areas of computing right now. Just last week, Facebook explained how it was using machine learning and artificial intelligence to understand the contents of photos that are uploaded to the site. While that may sound creepy, Facebook's intention is to make it easier for users to search for and dig up photos.


As it pertains to discriminatory advertising, Facebook is testing new technology that leverages machine learning to help identify ads that offer housing, employment, or credit opportunities, which are the same types of ads that advertising stakeholders told Facebook they're concerned about.

"This will allow us to more quickly provide notices and educational information to advertisers—and more quickly respond to violations of our policy," Facebook said in a blog post.

To that end, Facebook has made two specific changes:
  • Disapproving ads offering housing, employment or credit opportunities that use our multicultural affinity segments: When an advertiser attempts to show an ad that we identify as offering a housing, employment or credit opportunity and either includes or excludes our multicultural advertising segments—which consist of people interested in seeing content related to the African American, Asian American and US Hispanic communities—we will disapprove the ad.
  • Requiring self-certification: When an advertiser attempts to show an ad that we identify as offering a housing, employment or credit opportunity and uses any other audience segment on Facebook, we will show the advertiser information about our updated anti-discrimination policy. We will then require the advertiser to certify that it is complying with that policy and with applicable anti-discrimination laws.
These stronger enforcement tools are in addition to Facebook's recently updated policies, which are now more restrictive.

What Facebook is doing here is not an easy undertaking. There is a mountain of data accessible on Facebook, and advertisers use that data to target certain demographics. Where this became a pressing issue is when last fall there were ad campaigns that included or excluded certain users based on their ethnic affinity. By allowing those ads, Facebook was in violation of anti-discrimination laws, including the Fair Housing Act of 1968.

Since starting to implement changes last fall, Facebook says it's heard from several organizations interested in working with the social network on ways to use advertising to promote inclusion and opportunity for underserved communities while simultaneously protecting against discriminatory uses.