A study into ISIS propaganda on Facebook has found that the social media platform is failing to adequately deal with the terror group's shared content. Moderation has always been a concern for the social media giant and therefore, it is not a total surprise that extremist content is still being propagated on the platform.
Radicalization is one of the biggest threats to democratic societies and Facebook has long been accused of profiting from divisive content. The social media platform has been facing severe criticism for tolerating large-scale divisive campaigns by political groups, often involving misinformation, to sway voters during elections and on other key political issues. This has resulted in user boycotts as well as advertiser boycotts. From the company's perspective, Facebook CEO, Mark Zuckerberg, has been reluctant to take any serious measures to regulate the content on the platform, citing 'free speech' as the reason. However, following the recent protests and the massive advertiser boycott it would seem that Facebook is now making some changes.
Findings from a recent study suggest ISIS propaganda on Facebook is still a worrying concern. The research was conducted by the London-based think tank, the Institute for Strategic Dialogue (ISD). According to the study, researchers discovered that although Facebook removes almost 99 percent of terrorism-related posts, a large number of accounts that propagate pro-ISIS content have thrived by managing to evade Facebook's filters. The study focused on one such network, branded as Fuouaris Upload, and how it has created hundreds of accounts and reaches tens of thousands of users. The study claims that this is just a single network within a cluster of interconnected networks across Facebook.
Pro-ISIS Accounts Evading Facebook's Algorithms
Conducted during a three-month period between April and July, the study discovered 288 accounts in the Fuouaris Upload network with 90 of them controlled by a single user. The researchers found six specific ways in which these accounts were evading Facebook's automated and manual content moderation. First is by account hijacking where Fuouaris Upload members seem to be using a loophole in Facebook's account security protocol to intercept password reset text messages sent by the platform. Tutorial videos on how to do this are understood to have been shared between network members.
Content masking to make extremist content look like content shared by trusted news sources, such as the BBC and France24, is another way accounts are avoiding detection by Facebook. Link sharing, that is, sharing links in a decentralized manner through comment threads has also proved to be another popular tactic to spread ISIS content. Hashtag hijacking is yet another trick that is being used and this involves creating hashtags that resonate with an ongoing issue, for example, Black Lives Matter. These hashtags are used for raiding comment threads and masking extremist content. In addition, the pro-ISIS account members are also using broken texts, as well as specialized fonts, to evade Facebook's text analysis moderation.
While it is a relief that Facebook is managing to detect a majority of terrorist content, the existence of loopholes that can be exploited to spread harmful content is concerning. Hopefully, Facebook will act fast and fix these issues before the upcoming US presidential election, since the same loopholes could also be used by political groups, looking to polarize popular debate.
Source: ISD