In a reply to The Wall Street Journal’s report on Facebook’s underscoring AI, Facebook wrote a blog post explaining how well they’ve trained the AI on determining and removing hate speech.
The company’s Integrity VP countered with few statistics detailing how their hate speech handling AI has evolved over the years and is now able to accurately determine and tackle the hate speech prevailing on the platform.
Facebook AI’s Efforts on Countering Hate Speech
For a long, Facebook has been criticized for various reasons, with growing hate speech being one among them. Though the platform claims in handling hate speech finely, a report from The Wall Street Journal recently said Facebook’s AI has failed in this job.
Now, quashing the WSJ’s report, Facebook’s Integrity VP Guy Rosen explained in a blog post that,
“We don’t want to see hate on our platform, nor do our users or advertisers, and we are transparent about our work to remove it.”
Further, “What these documents demonstrate is that our integrity work is a multi-year journey. While we will never be perfect, our teams continually work to develop our systems, identify issues and build solutions.”
He went on to explain that Facebook doesn’t measure success in this field by the number of content removals it does, but by observing how much hate speech is prevailing on its platform. Thus, he affirmed that it’s a continuous journey, and they’re developing the AI to be better in it.
Now, Facebook’s AI is able to handle over 97% of hate speech reports on the platform, up from 23.6% earlier, with the rest being handled by people.
Any hate speech their AI has detected will be removed once the bar is reached, or made less visible to others and groups. The AI has grown significantly that, hate speech is now being viewed by only five people per every 10,000 on its platform.
Other Trending News:- News