LEA Working Group Insights on Online Hate Speech
In its fourth edition, the CC-DRIVER LEA Working Group (WG) has shared challenges, resources and best practices on tackling online hate speech. This virtual meeting has been organised by Valencia Local Police on Friday, 1 October 2021. Five participants from five different countries (Finland, Germany, Greece, Spain and Switzerland) joined the gathering.
It was observed that there is no common legal definition of online hate speech among EU Member States and the legal landscape as to prohibiting content clearly differs. It was pointed out in the LEA WG meeting that this constitutes a further difficulty to measuring hate speech content and its current trends. In this sense, the EU Commission initiative for extending the list of EU crimes to all forms of hate crime and hate speech – which is scheduled for the fourth quarter of 2021 – could help LEAs not only in preventing, but also in predicting and measuring hate speech content. Underreporting hate speech was also discussed as another challenge LEAs face when trying to determine the real dimension of this issue, especially when a great part of online hate speech takes place in private groups or communication platforms such as WhatsApp or Telegram. It was also noted that standardisation at a European level, possibly in the form of government statistics on different types of illegal online content, including both hate speech and online harassment would further help LEAs in grasping the issue and thus generating the most appropriate strategies to counter it.
The LEA Working Group also discussed the role of LEAs in monitoring Internet and social media and agreed that, unless for specific investigations, as well as specialised LEA departments, it is not considered a “core business” for them. As such, the use of Artifical Intelligence (AI) by LEAs is currently not a priority, at least as far as removing hate speech from social media is concerned; a task that usually does not involve LEAs, but rather general users or so-called “trusted flaggers”, both of which do not necessarily have to be an LEA. Moreover, the participating LEAs agreed that AI’s use for detecting online hate speech would be challenging because it should not undermine freedom of expression and one would have to find the right balance. It is worth noting that the EU proposal for an AI Act may help LEAs by providing a framework for the competences and procedures on the use of AI for this purpose. As a final note, the PROPHETS H2020 project was mentioned, where a mechanism has been developed for detecting terrorism-related hate speech, providing satisfactory results.
Regarding best practices on countering hate speech online, the EU Code of Conduct has been proven to be a valuable tool to prevent and counter the spread of illegal hate speech online by incentivising proper moderation. In its sixth evaluation, which has been recently published (7 October 2021), it states that “while the average of notifications reviewed within 24 hours remains high (81%), it has decreased compared to 2020 (90.4%). At 62.5% the average removal rate was also lower than in 2019 and 2020”. On the other hand, a common view in the LEA WG was that law enforcement agencies usually are not effective in preventing and mitigating online hate speech due to the possible reluctance of at-risk individuals to contact and communicate with police forces. Instead, a multi-actor approach where civil society organisations, non-government organisations or charities are direct communicators with the at-risk individuals (coordinating efforts with LEAs) is advised, given the positive results LEAs (for instance, the Police Service of Northern Ireland) have achieved with this approach. Lastly, the key role of the No Hate Speech Movement was mentioned, which is a youth campaign led by the Council of Europe’s Youth Department seeking to mobilise young people to combat hate speech and promote human rights online.On its website, it provides information about the campaign and the resources developed to prevent, counter and produce alternative narratives to hate speech.
Descargue esta entrada de blog en español: