Ad discrimination is a pretty serious issue on Facebook, and when those in charge of the social network found out that some advertisers were using “ethnic affinity” ads for housing, employment or credit, they decided to fight the issue using algorithms. While the first anti-discrimination methods were moderately successful, Facebook wants to take things a step further by educating advertisers that they are not allowed to discriminate users based on their gender, race, age, color, family status, ethnicity, religion or medical condition. Furthermore, the company will leverage machine learning in order to flag all ads relating to employment, credit or housing, and it will disapprove these ads if they do not comply with the recent anti-discrimination guidelines.
When a user starts to create an ad, he will be reminded of these new rules and will be asked to double-check if the ad in question is not in any way discriminatory. We’ll just have to wait and see if these new methods will harm advertisers in the long run, as algorithms are known to skip a few beats every now and again, meaning that there’s always room for error. Leaving the anti-discrimination checks in the hands of the advertisers themselves doesn’t really sound like the best solution, but it is definitely in their best interest to follow the new guidelines if they want their ads to appear at all.
