Facebook has found itself in another mess it could have perhaps done without. Trouble started when its automated advertisement system used one of Instagram user’s posts to publicize its offering to other users of the platform. Nothing wrong with that except that the post contained the term, ‘I will rape you before I kill you.’
The hate post again belonged to one Olivia Solon, a reporter with Guardian which she received some time back from someone who she said she’d prefer to identify as an idiot than anything else. However, Instagram used the screenshot of that post as publicity material to highlight Instagram to Solon’s own sister.
Instagram meanwhile has been quick on the damage control act, saying the above was not part of a paid promotion drive, besides also apologizing for it as well. Instagram too joined in saying they would be introducing enough changes to its algorithms to ensure such things don’t recur again. Facebook’s chief operating officer Sheryl Sandberg said such targeted ads in future would be allowed to go live only after they have been previewed by humans.
Instagram often uses other user’s posts to highlight its services to dormant members, or those who haven’t been online in the recent past. Not surprisingly, such posts are used that have evoked considerable amount of interest from other users.
In this case, the post chosen from Solon’s timeline is found to have three likes while about a dozen comments were made.
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
Those figures likely could have made a match as per parameters set by the auto ad algorithm, enough to make use of the same as advertisements material.
However, what seems to have gone wrong is that the program failed to make out what the actual content of the post is. And as it would have been, it is a highly derogatory word that ended up being part of the Instagram promotional material, perhaps the worst that any woman can ever receive.
The goof-up also seems surprising given that the technology that allows a computer program to actually decipher the meaning of words used in a sentence already exists. Maybe Facebook feared using the same would have evoked privacy issues.
Facebook had also been hogging the limelight just last week thanks to another issue, that of allowing advertisers to promote their posts to Jew-hating community.