Facebook's vice president of integrity, Guy Rosen, wrote in a blog post on Oct. 17 that the number of hate speech on the social media platform had decreased by almost 50% over the past three years.
Rosen added that the narrative that the technology that they use to fight hate speech on the platform is inadequate, and that they deliberately misrepresent their progress was not true at all.
Facebook's Hate Speech Number Decreased
Rosen wrote in the Facebook blog post that they do not want to see hate on their platform and they do not want their users or advertisers to see them either. They are transparent about their work and effort to remove all of it, according to The Verge.
Rosen added that what the documents show is that their integrity work is a journey that takes years. While they admit that their system isn't perfect, their teams are constantly working to create systems to fight hate speech, identify the issues, and create effective solutions.
The post appeared to be in response to an article released by The Wall Street Journal on Oct. 17, which said that the company's employees tasked with keeping offensive posts off of the social media platform do not believe that Facebook is able to effectively screen for it.
Facebook Employees Speak Up
The Wall Street Journal report states that the internal documents show that in 2019, Facebook decreased the time that reviewers focused on hate speech complaints, and they made adjustments that reduced the total number of complaints.
With that strategy, it helped create the idea that the platform's artificial intelligence had been successful in enforcing its rules than it actually was.
In March, several Facebook employees found that the company's automated systems were removing posts that created 3% and 5% of the views of hate speech on the platform and less than 1% of all content that was in violation of its guidelines against violence.
However, Rosen stated that focusing on content removals alone was not the correct way to fight hate speech. He said that the technology to remove the posts is just one method that the company uses to fight it. He added that they need to be confident that something is hate speech before they take it off the site.
Rosen added that the company believes focusing on the prevalence of hate speech users actually see on the platform is a more vital measure.
Rosen claimed that for every 10,000 views of a post on the platform, around five of them are hate speech. He wrote that prevalence tells them what violating content users see because they missed it. It is how they evaluate their progress, as it gives the most complete picture.
But the documents that The Wall Street Journal obtained showed pieces of content that were able to evade the platform's detection, including videos of vehicles crashes that showed people with injuries and threats against children.
The Wall Street Journal has released several reports about the social media site based on the documents given by Frances Haugen, Facebook's whistleblower.
Haugen testified before Congress and said that Facebook was aware of the negative impact of its platform on users, especially on teenagers.
Related Article : Is Facebook Doing Enough To Remove Hate Speech? Austrian Court Thinks Not, But This Ruling Should Change Things
This article is owned by Tech Times
Writtn by Sophie Webster
