On February 28th, shortly after the Russian invasion of Ukraine, Twitter announced the expansion of its labelling policy for “Russia state-affiliated media”, in order to address disinformation in favour of the Russian government. While this ‘soft’ approach does not include the removal of content, it entails issues for the freedom of expression and information. This article addresses the consequences of this labelling policy for the range and impact of accounts labelled “Russia state-affiliated media” during the Ukrainian war. Using an iterative detection method, a total of 90 accounts of both media outlets and individual journalists with this label were identified. The analysis of these accounts’ information and timeline, as well as the comparison of the impact of their tweets before and after February 28th with an ARIMA model, has revealed that this policy, despite its limited scope, lead to a significant reduction in the impact of the sampled tweets. These results provide empirical evidence to guide critical reflection on this content moderation policy.

In recent decades, many sectors of our society have been digitized, and much of our life has moved to cyberspace, especially in terms of entertainment. Users meet, relate, and cooperate in the new public space that is the internet and form digital communities. Video games play a leading role in the formation of such communities. However, these communities also present antisocial behaviors, ranging from disruptive actions to harassment and hate speech. Such behaviors, encompassed under the umbrella term toxicity, are a major concern for both users and those in charge of moderating these spaces. This article focuses on toxicity in today’s leading online video game League of Legends. Three hundred twenty-eight matches were reviewed using a system of two judges to study the prevalence of these problematic behaviors. We find that 70% of matches were affected by disruptive behavior. Nevertheless, only 10.9% of the analyzed matches were exclusively affected by downright harmful behavior. In our view, the results have relevant implications for content moderation policy that are also addressed in this paper.

Disinformation has been described as a threat to political discourse and public health. Even if this presumption is questionable, instruments such as criminal law or soft law have been utilised to tackle this phenomenon. Recently, technological solutions aiming to detect and remove false information, among other illicit content, have also been developed. These artificial intelligence (AI) tools have been criticised for being incapable of understanding the context in which content is shared on social media, thus causing the removal of posts that are protected by freedom of expression. However, in this short contribution, we argue that further problems arise, mostly in relation to the concepts that developers utilise to programme these systems. The Twitter policy on state-affiliated media labelling is a good example of how social media can use AI to affect accounts by relying on a questionable definition of disinformation.