Moderating a Facebook gardening group in western New York City is not without its challenges. There are complaints of woolly bugs, bad weather, and novice members who insist on using dish detergent on their plants.
And then there’s the word “hoe”.
Facebook’s algorithms sometimes flag this particular word as “violating community standards,” apparently referring to a different word, a word without an “e” at the end that is nonetheless often misspelled as a gardening tool.
Normally, Facebook’s automated systems flag posts containing offending material and delete them. But if members of a group – or worse, admins – break the rules too often, the entire group can be shut down.
Elizabeth Licata, one of the group’s moderators, expressed concern. After all, the group, WNY Gardeners, has over 7,500 members who use it for gardening tips and advice. It was particularly popular during the pandemic when many people confined to the house began to garden for the first time.
A hoe of any other name can be a rake, harrow, or tiller. But Licata wasn’t about to ban the word from the group, or try to remove every instance. When a band member commented “Push pull hoe! “On a post asking” your most loved and indispensable weeding tool, “Facebook sent a notification that said” We have reviewed this comment and found it to be against our harassment and harassment standards. intimidation”.
Facebook uses both human moderators and artificial intelligence to eliminate content that goes against its rules. In this case, a human would probably have known that a hoe in a gardening group is probably not a case of harassment or bullying. But AI isn’t always good at context and nuances of language.
There’s a lot missing, too – users often complain about reporting violent or abusive language, and Facebook’s policies say it’s not in violation of its community’s standards. Misinformation about vaccines and elections is a long-standing and well-documented problem for the social media company. On the other hand, there are groups like Licata’s that are caught up in overzealous algorithms.
“And so I contacted Facebook, which was pointless. How do you do that? “She said.” You know, I said it was a gardening group, a hoe is a gardening tool. ”
Licata said she never heard from someone and from Facebook, and found it futile to navigate the social network’s investigation system and ways to try to set the record straight .
Contacted by The Associated Press, a Facebook representative said in an email this week that the company had found the group and fixed the application errors. It also implemented an additional check, which means that someone – a real person – will check the offending messages before the group is considered for deletion. The company didn’t want to say if other gardening groups had similar issues. (In January, Facebook reported in error British Plymouth Hoe landmark as offensive and then apologized, according to The Guardian.)
“We plan to develop better customer support for our products and provide the public with even more information about our policies and how we enforce them,” Facebook said in a statement in response to Licata’s complaints.
Then something else happened. Licata received a notification that Facebook was automatically disabling comments on a post due to “possible violence, incitement or hate in multiple comments.”
Offensive comments included “Kill them all. Drown them in soapy water ”, and“ Japanese beetles are fools ”.