Incident 283: Facebook’s Automated Content Moderation Tool Flagged a Post Containing Parts of the Declaration of Independence as Hate Speech by Mistake

Description: Facebook’s content moderation algorithm was acknowledged by the company to have flagged excerpts of the Declaration of Independence posted by a small newspaper in Texas as hate speech by mistake.
Alleged: Facebook developed and deployed an AI system, which harmed The Vindicator.

Suggested citation format

Lam, Khoa. (2018-07-02) Incident Number 283. in Lam, K. (ed.) Artificial Intelligence Incident Database. Responsible AI Collaborative.

Incident Stats

Incident ID
Report Count
Incident Date
Khoa Lam


New ReportNew ReportDiscoverDiscover

Incidents Reports

  • Facebook has apologized for taking down a post containing excerpts of the Declaration of Independence, saying that it was mistakenly flagged as hate speech.
  • In addition to apologizing, Facebook has restored the post.
  • This is just the latest high-profile example of Facebook's hate speech algorithm making a mistake.

In yet another viral case of Facebook struggling to police hate speech on its platform, parts of the Declaration of Independence posted by a newspaper in Texas were taken down earlier this week after the social media giant flagged the excerpts as hate speech.

The post has since been restored and Facebook has apologized.

The Vindicator, a small community newspaper in Liberty County, Texas, started posting excerpts of the Declaration of Independence earlier this week leading up to the Fourth of July. While the newspaper was able to post the majority of the Declaration of Independence without any issue, one post contained the phrase "Indian Savages," which, out of context would appear to violate Facebook's community standards.

Here is the entire sentence:

“He has excited domestic insurrections amongst us, and has endeavoured to bring on the inhabitants of our frontiers, the merciless Indian Savages, whose known rule of warfare, is an undistinguished destruction of all ages, sexes and conditions.”

Facebook's automated systems could have detected the use of "Indian Savages," and triggered the platform to take the post down, the company said. But after The Vindicator's editors published a story about it on July 3 and notified Facebook, the company restored the post and apologized.

“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," a Facebook spokesperson told Business Insider.

This is just the latest high profile example of a case of Facebook's hate speech algorithm making a mistake. Facebook uses a combination of humans and automation to review posts, and even though the company said it is aware of the complex issues surrounding hate speech — like intent and context — sometimes errors are made.

Facebook has said it plans to hire thousands more human reviewers, but this incident shows how the social media giant is still working out how best to remove hate speech without censoring legitimate expression.

Facebook has apologized for flagging parts of the Declaration of Independence as hate speech