Online extremism and hate speech on the internet are growing global concerns. In 2015, the EU signed a code of conduct with social media companies including Facebook, Google and Twitter to effectively regulate hate speech on the internet. The code, amongst other measures, discussed stricter sanctions on intermediaries (social media companies) in the form of a ‘notice and takedown’ regime, a practice which has been criticised for effectively creating a ‘chilling’ effect and leading to over-censorship.
While this system is still in place, social media companies are attempting to adopt alternative regulatory methods. If companies could ensure that they routinely track their websites for illegal content, before government notices are issued, this could save them time and money. This post will attempt to offer some insight into alternative modes of regulation used by social media companies.
YouTube Heroes – Content Regulation by Users
YouTube Heroes was launched in September, 2016 with the aim of efficiently regulating content. Under this initiative, YouTube users are allowed to ‘mass-flag’ content that goes against the Community Guidelines. The Community Guidelines specifically prohibit instances of hate speech. As per the Guidelines, content that “promotes violence or hatred against individuals based on certain attributes would amount to hate speech”. These ‘attributes’ include but are not limited to race, gender and religion.
‘Mass-flagging’ is just one of the many tools available to a YouTube Hero. The system is based on points and ranks, with users generating points for helping translate videos and for flagging inappropriate content. As they climb up the ranking system, users become privy to exclusive deals, like the ability to directly contact YouTube staff. ‘Mass-flagging’ is in essence the same as flagging a video, an option that YouTube already offered. However, the incentive of gaining access to private moderator forums and YouTube staff could lead to users flagging videos for extraneous reasons. While ‘mass-flagged’ videos are reviewed by YouTube moderators before being taken down, the initiative has still raised concerns.
It has been criticised for giving free rein to users, who may flag content because of personal biases, leading to ‘harassment campaigns’. Popular YouTube users have panned YouTube heroes, apprehending the possibility of their videos being targeted by ‘mobs’. Despite the review system in place, users have also expressed doubts about YouTube’s ability to accurately take down flagged content. Since the initiative is in its testing stage, it is difficult to determine what its outcome could be.
Facebook’s Online Civil Courage Initiative – Counter Speech
Governmental authorities across the world have been attempting to curb hate speech and online extremism in myriad ways. For instance, in November, 2015, an investigation involving one of Facebook’s European Managing Directors was launched. The Managing Director was accused of letting Facebook host hate speech. As the investigation drew to an end, Facebook representatives were not implicated. However, this investigation marked an increase in international pressure to effectively deal with hate speech.
Due to growing pressure from governmental authorities, Facebook began to ‘outsource’ content removal. In January of 2016, a German company called ‘Arvato’, was delegated the task of reviewing and taking down reported content, along with Facebook’s Community Operations Team. There is limited public information on the terms of service or rules Arvato is bound by. In the absence of any such information, ‘outsourcing’ could contribute to a private censorship regime. With no public guidelines in place, the outsourcing process is not transparent or accountable.
Additionally, Facebook has been working with other private bodies to regulate content online. Early in 2016, Facebook, in partnership with several NGOs, launched the Online Civil Courage Initiative (OCCI) to combat online extremism with counter-speech. COO Sheryl Sandberg said that ‘censorship’ would not put an end to hate speech and that counter-speech would be a far more effective mode of regulation. Under this initiative, civil societies and NGO’s are ‘rewarded’ with “ad credits, marketing resources, and strategic support” for countering speech online.
It is pertinent to note that the Information Pack on Counter Speech Engagement is the only set of guidelines made public by OCCI. These guidelines provide information to plan a counter speech campaign. An interesting aspect of the information pack is the section on ‘Responding and Engaging during a campaign’. Under this section, comments are categorised as ‘supportive, negative, constructive, antagonistic’. A table suggests how different categories of comments should be ‘engaged with’. Surprisingly, ‘antagonistic’ comments should be ‘ignored, hidden or deleted’. The information pack does not attempt to define any of the above categories. These vaguely worded guidelines could lead to confusion amongst NGOs. While studies have shown that counter-speech might be the most effective way to deal with online extremism, OCCI would have to make major changes to reach the goals of the counter-speech movement.
In October 2016, Facebook has reportedly come under the radar again. A German Federal Minister has stated that Facebook was still not effectively dealing with hate speech targeted at refugees and another investigation might be in the pipeline.
It is yet to be seen whether the alternative regulatory methods adopted by social media companies will effectively deal with hate speech and online extremism.
It is important to note that social media companies are ‘outsourcing’ internal regulation to private bodies or users (YouTube Heroes, Arvato and OCCI). These private bodies might amplify the problems being faced by the intermediary liability system, which could lead to ‘over-censorship’. The system has been criticised for its ‘notice and takedown’ regime. Non-compliance of these takedown orders would attract strict sanctions. Fear of these sanctions could lead intermediaries to takedown content which could be in grey areas, but are not illegal.
However, under the internal regulation method, social media companies will continue to function under the fear of state pressure. Private bodies like Arvato and NGOs in affiliation with OCCI will also regulate content, with the incentive of receiving ‘advertisement credit’ and ‘points’. This could lead to over-reporting for the sake of incentives. Coupled with pressure from the state, this might lead to a ‘chilling’ effect.
In addition, some of these private bodies do not operate in a transparent manner. For instance, providing public information on Arvato’s content regulation activities and the guidelines they are bound by would help create a far more accountable system. Further, the OCCI needs to have clearer, well-defined policies to fulfill the objectives of disseminating counter-speech.