Another day, another scandal involving digital ad platforms. Today’s entry comes from Instagram, which made an advertisement of a Guardian reporter’s post that documented a threatening email she received. The message said, among other things, “I will rape you before I kill you, you filthy whore!”
The reporter, Olivia Solon, discovered the ad when it was pushed to her sister. It seems to have automatically picked that Instagram post because it received a lot of engagement–sympathetic comments and a few likes. The ad highlighted the threatening post along with the message “See Olivia Solon’s photos and posts from friends on Instagram”–quite the contrast. Instagram apologized for letting it happen.
Instagram is using one of my most “engaging” posts to advertise its service to others on Facebook ???? pic.twitter.com/lyEBHQXMfa
— Olivia Solon (@oliviasolon) September 21, 2017
Currently platforms like Instagram, Facebook, Google, and Twitter are being scrutinized about the workings of their ad products. They use algorithms and other technology to figure out the best way to target to people, but so much can go wrong in the interim. For example, ProPublica reporters discovered that they could target Facebook ads to a group whose interest was “jew hater.”
This latest Instagram example shows the pitfalls of using technology to automatically make decisions. If it’s unable to deduce that “I will rape you” is not appropriate for advertising purposes, perhaps a human should be included in the mix.