Facebook has detailed how it is using artificial intelligence to detect and remove hate speech and misinformation related to the COVID-19 pandemic.
The social media giant notes that the same piece of misinformation can be spread in several forms like an image modified with a few pixels cropped or manipulated in some other way. This could either be done accidentally or deliberately to avoid detection.
“AI is a crucial tool to address these challenges and prevent the spread of misinformation, because it allows us to leverage and scale the work of the independent fact-checkers who review content on our services,” the platform stated in a blog post.
Facebook has partnered with over 60 fact-checking organizations around the world to review content in more than 50 languages. It says that since the pandemic started, it has used its current AI systems and created new ones to review COVID-19 posts.
The fact-checking organizations have flagged posts that contain misinformation, and have detected copies where other users try to share the same content.
Since Facebook has banned ads for face masks and some other products, it has trained its AI to detect listings that may have made it through its systems by modifying the content. The social media says that in most cases, it has been able to detect most ads before anyone has flagged them.
Facebook says that it put warning labels on around 50 million posts related to COVID-19 in April alone. Since March, it has removed more than 2.5 million posts that tried to sell masks, hand sanitizer and fake COVID-19 testing kits.
Further, the platform is also taking on hate memes since they can often go undetected as hate speech by AI systems. It has released a ‘Hateful Memes’ data set that contains 10,000+ examples to help systems better understand them.
