Preventing user-generated hate speech

Contrary to countries like the United States, where the First Amendment prevents the regulation of hate speech, in Romania the law prohibits quite clearly the discriminatory behaviors generally associated with hates speech. This is visible, for example, in the fact that all the sites analyzed in this study transposed the legislation into their guidelines prohibiting that type of messages, which however does not prevent online hate speech from happening.

Therefore the question is not whether hateful comments on websites should be filtered, rather why are they still there despite the legislation and what could be done to prevent the phenomena.

In my view there could be three solutions to user-generated hate speech: a) the separation of the readers’ comments from the main page; b) amendments to the legislation to make newspapers responsible for their electronic pages, and c) self-regulation by newspapers. The last would be the optimal solution as it would prevent state interference into the media system.

CEUeTDCollection

The legislative option would require the government to elaborate a special law on online media (where it would clarify the question of responsibility for user-generated content on mainstream sites differentiated from dedicated UGC sites¹⁷⁰), or the participation management requirements for the service providers. The modification of the existing legislation could add content monitoring function with regards to discrimination to the CNCD also extended to websites, which would allow for the institution to issue take down notices for UGHS - but this second solution would not prevent the initial posting of such messages.

Neither of these solutions is desirable, as they would increase the potential for state interference and abuse in the media, especially in presence of authoritarian government. The simplest statutory solution would be a rigorous enforcement of the legislation already in place, and the prosecution of hate speech cases in comment sections, which combined could act as a deterrent. However, this last solution could have a chilling effect on legitimate comments, and it is again open for state abuses.

In my opinion the optimal solution would be self-regulation, a public trust approach as described by McQuail.¹⁷¹ Media organizations could agree to an ethics code for audience participation which would also contain good practice recommendations regarding moderation and a commitment by the participating newspapers/sites towards moderating comments, based on a commonly agreed set of guidelines similar to the codebook in appendix 2. As it happens with the codebook, such guidelines could also incorporate elements of the legislation, which in the current approach is totally ineffective. Having a document similar to the codebook would ensure that the terms “discriminatory”, “instigating”, “racist”, and

“hateful” already present despite being prohibited in the guidelines of the sites finally have a shared meaning. The codebook itself could also be published in the TOS to make users aware of the actual meaning and extent of the prohibited categories.

170 such as the Youtube video sharing platform

171 McQuail,Accountability of Media

CEUeTDCollection

A major obstacle in implementing such a solution is the lack of a joint self-regulatory body in Romania and the financial costs of moderation, which imposes a burden on media companies.

On the one hand the moderation itself is costly, but there is also the indirect cost of potentially reduced advertisement revenues. As the content analysis shows, 37.99% of comments contain hate speech: eliminating them would cut in almost half the levels of user participation.

Additionally if the hate comments would not be published the comments reacting to them would also disappear, resulting not only in reduced participation but also in reduced advertising revenues as visitors on the site would spend less time reading comments or reacting to them. On the other hand, taking the example of the article with the highest proportion of UGHS (78.64 percent), the 674 comments currently posted on the site (even disregarding multiple contributions) represent still an insignificant fraction of the 28514 views of the article, who presumably went there to read the editorial and not the user-generated content.

On the other hand the unregulated and unrestricted nature of the comments creates a responsibility-free space on the online newspaper which is not present in its offline counterpart, and that can also be used for manipulating the public’s perception about the issues presented in the articles or even for the intimidation of a given group – something media are not able to do in the offline or even in the professional areas of the online world due to the threat of possible legal actions. Media organizations and journalists both offline and online are constrained by professional guidelines, ethical rules, laws and other similar formal or informal regulations regarding content; breaching these has legal and moral consequences on their professional reputation. On the other hand, as the data presented earlier shows, there are no such limits in the comments. Views that cannot be published due to societal norms, laws or ethics in an article can be published in the comment sections of an article on the same topic, while still benefiting from the same audience. In many cases such as

CEUeTDCollection

the two topics with the highest amount of hate-speech: the territorial reorganization and Roma criminality, the amount of hate comments and their nature is evident at a first glance. It is unlikely that the administrators of the sites were not aware of having such content displayed on their pages. Therefore, one has to wonder about their motivations for allowing it even in clear infringement of their own guidelines. Returning to the article presented at the beginning in the thesis the fact that the site used an inflammatory title and allowed such high amount of extremely hateful and violent comments might also be intended as an intimidation of the minority.