Facebook has issued an apology to a local Texas publisher after removing a post containing text from the Declaration of Independence after it was flagged as “hate speech”.
The Vindicator shared a series of excerpts from the Declaration to its Facebook page in advance of the Fourth of July. Managing editor Casey Stinnett said the 10th installation, consisting of paragraphs 27 to 31, did not appear in the post and the publication instead received a notice from Facebook saying it “goes against our standards on hate speech."
“We define hate speech as a direct attack on people based on what we call protected characteristics,” said the company.
“We define attack as violent or dehumanizing speech, statements of inferiority, or calls for exclusion or segregation.”
Facebook currently has 15,000 human moderators trained to detect hate speech and has plans to hire another 5,000 by the end of the year, but they only respond to posts Facebook users have flagged. To combat the billions of posts generated each day, the company has implemented an automated action that uses an artificial intelligence algorithm for deciphering and blocking hate speech.
AI censorship has become an increasingly contentious issue as society navigates the waters of fake news. For starters, understanding the meaning of language goes beyond strictly pinpointing certain buzzwords, such as the context of this historical document. This extends beyond the written word and into the world of video as well, as DeepFakes could prove too difficult for machine learning to accurately flag and pinpoint.
The publication says they are unsure what prompted the filtering program, but believe it could have been a phrase that includes the following:
“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.”