Can AI replace humans? That’s the question that many tech companies want to know. In some ways, computers are already replacing humans, as is the case with Facebook advertising. Facebook has created algorithms that create targeted advertising categories and approve ads that are placed on the site. But often, as is currently the case, computers can go awry.
Facebook has recently come under fire for its advertising ways. Most recently, the company has been targeted for allowing advertisers to create ads aimed at ‘anti-Jewish’ populations. The investigative journalism network, ProPublica, attempted to post fake advertisements that targeted anti-semitic populations.
What journalists discovered has made Facebook a news target this morning.
The ProPublica Trap
ProPublica asked investigative journalists to try and post anti-semitic ads on Facebook through the company’s (mostly automated) advertising network. Three ads were successfully posted and filed under the category ‘History of Why Jews Ruined the World.’ The ads were approved in 15 minutes through Facebook’s algorithms.
Shortly after Facebook admins caught wind of this issue, the posts were removed along with the category that permitted the ads to be posted. Facebook has since addressed press stating that its algorithm created the category and approved the ads, and that the company realizes it has work to do in those regards. But has a lot of damage already been done?
The Russia Problem
Facebook is currently being scrutinized by the US Government for potentially allowing a Russian-based group to post $1,000 worth of Facebook ads promoting various political messages. The ads exploited the holes in Facebook’s advertising algorithm and allowed the group to potentially skew the latest US presidency vote.
Facebook is currently working with government officials to find the root cause of this problem and to determine how much influence the ads had over the latest US election. Further, there is some evidence to suggest that Russian groups set up numerous Facebook groups and used those groups to create anti-immigration and other rallies in the US and elsewhere. The fact remains that it is hard to tell who is behind a Facebook group, exactly.
Back to the Human Quotient?
Does this bring Facebook back to human moderators when it comes to advertisements and advertising categories? Probably not. It would cost the social network too much to staff all of those advertising submissions and create new categories for ad placement based on observations. At the same time, the company has to find a way to stop its algorithm from doing two very bad things.
The first is allowing hate ads to be posted on the network - and allowing those ad categories to even exist. The second is allowing fake accounts to be created. It will be very hard for Facebook to crack down on those accounts, though. First, the company has to figure out who is behind each account (hard to track down when people use fake information), and then close those offending accounts.
In other words, Facebook has its work cut out. At some point, though, someone has to reel in the robots.