Facebook Flags Declaration of Independence as Hate Speech

Facebook removed a post with sections of the Declaration of Independence because it was "hate speech." It's not clear if the screw-up was a human error or a failure of an automated filter.

The post came from Texas newspaper the Liberty County Vindicator, which posted the Declaration of Independence as a series of scheduled posts to mark July 4th. One of the posts was flagged and removed by Facebook while it was still in the scheduled queue, yet to go live. The newspaper received a message saying the deletion was because the post "goes against our standards on hate speech."

Indians Reference May Be Issue

The post in question contained the 27th through 31st paragraphs of the Declaration of Independence, which contain part of the long list of charges of misconduct by King George III. While the deletion notice didn't say exactly which words caused the problem, the newspaper believes the most likely culprit was a reference to inciting "the merciless Indian Savages." (Source: thevindicator.com)

The Vindicator staff believe the post must have been removed automatically, figuring no human moderator would make such a mistake, though there doesn't seem to be any proof either way.

Facebook Admits Mistake

Facebook later restored the post and messaged the newspaper to say "It looks like we made a mistake and removed something you posted on Facebook that didn't go against our Community Standards. We want to apologize and let you know that we've restored your content and removed any blocks on your account related to this incorrect action."

Facebook later told CNN that "the post was removed by mistake and restored as soon as we looked into it. We process millions of reports each week, and sometimes we get things wrong." (Source: cnn.com)

Assuming it was an algorithm that removed the post, it's a sign of the difficulties of computers analyzing context when reading words and phrases literally, particularly historical context.

What's Your Opinion?

Do you think this was human error or flawed automation? Could Facebook realistically improve its filters to better judge context? Would it be realistic to have algorithms spot potential hate speech but then have human moderators check before taking it down?