[ad_1]
In regards to measuring its progress in tackling hate speech, Facebook announced today that, “for the first time,” it includes the prevalence of hate speech on the social network, as part of its quarterly report on application of community standards.
Prevalence, he explains, estimates the percentage of times people see such content on that platform.
“We calculated the prevalence of hate speech through a sample of content viewed on Facebook and tagged when it violates our hate speech policies” and, as this type of speech “depends on cultural context and language, we send these representative samples to reviewers in different regions and languages, “explains Facebook.
Based on this methodology, “we estimate that the prevalence of hate speech from July to September 2020 was 0.10% to 0.11%”.
That is, “out of 10,000 views, 10 to 11 included hate speech,” the social network said.
In the third quarter, Facebook removed 22.1 million hate speech content and Instagram 6.5 million.
“We specifically measure how much malicious content can be seen on Facebook and Instagram, because the number of times the content is viewed is not distributed in the same way.”
In other words, a piece of content can go viral and be seen by many people in a short period of time, while other content can stay on the Internet longer and only be seen “by a small group”, continues the social network.
Facebook adds that defining hate speech is not a simple matter, as there are differing views on what constitutes that speech. History, language, religion and changing norms are important factors to consider in how the social network defines its policies.
“Based on input from a wide range of global experts and stakeholders, we define hate speech as anything that directly attacks people based on protected characteristics, including race, ethnicity, national origin, religious affiliation, sexual orientation, sex, gender, gender identity or disability or serious illness, “he explains.
“Our goal is to remove hate speech whenever” it is detected, but “we know we still have progress to make,” Facebook points out.
The social network has developed artificial intelligence (AI) tools and systems to proactively detect and remove hate speech.
“Because hate speech is so contextual, AI detection requires the ability” to look at posts holistically, with context, and in multiple languages, he points out.
However, advances in AI technologies allow us to increase the removal of this type of content on Facebook.
The social network says that when it began reporting its hate speech metrics in Q4 2017, the proactive detection rate was 23.6%.
“This means that of the hate speech we removed, 23.6% were detected before any user reported it,” he explained to Facebook.
“Today, we proactively detect 95%” of this content, which is removed.
Source link
