On Thursday, Facebook announced that their artificial intelligence software detects 94.7% of the hate speech that gets taken down from their website.
Facebook’s chief technology officer, Mike Schroepfer, said in a blog post that this number was 80.5% last year and just 24% in 2017.
Even though Facebook employs thousands of people to moderate their content and take down any fake posts, photos, or videos, humans alone cannot meet the challenge. Facebook, along with Twitter and TikTok, relies on artificial intelligence, where algorithms improve automatically through experience.
Schropfer said, “A central focus of Facebook’s AI efforts is deploying cutting-edge machine learning technology to protect people from harmful content.” He added, “Our goal is to spot hate speech, misinformation, and other forms of policy-violating content quickly and accurately, for every form of content, and for every language and community around the world.”
However, Facebook’s AI software still fails to detect some posts that break the rules, such as photos with underlying messages or posts with sarcasm and slang, in which case humans are required to detect and take down the posts.
Facebook said it is installed two new AI technologies to help with locating hate speech. One is “Reinforced Integrity Optimizer” which “learns from real online examples and metrics instead of an offline dataset” and the second is “Linformer” which allows Facebook to use “complex language understanding models that were previously too large and unwieldly to work at scale.”
Schroeper said they use these two AI technologies to analyze Facebook and Instagram content in different countries globally.
In October, Facebook said they were updating their hate speech policy to take down any content which denies or spreads misinformation about the Holocaust. However, India spreads a vast amount of hate on platforms such as Instagram, Facebook, and WhatsApp, mostly targeting Pakistanis and minorities in their own country, including Muslims.
In May, a top UN official voiced concerns over “increased hate speech and discrimination” in India against minorities, following the implementation of the Citizenship Amendment Act, under which India granted citizenship to all non-Muslim minorities who migrated from Afghanistan, Pakistan and Bangladesh till December 31, 2014.
The discrimination against Muslims is “contrary to India’s obligations under international human rights law, in particular on non-discrimination,” the UN official added. Additionally, he encouraged the Indian government to follow international standards of non-discrimination and to combat the rise of hate speech through messages of inclusion and respect, which the government has clearly failed to do.
Much of the hate speech India is spreading is not being taken down from these websites, which raises the question: why aren’t the same hate speech policies being implemented in India?
