Maja Brkan, Maastricht University Faculty of Law, has published Freedom of Expression and Artificial Intelligence: On Personalisation, Disinformation and (Lack Of) Horizontal Effect of the Charter. Here is the abstract.
The question whether the use of Artificial Intelligence can interfere with the freedom of expression and information has attracted a considerable amount of attention on both sides of the Atlantic. This paper contributes to this debate by questioning whether the Charter of Fundamental Rights of the EU can be considered as an appropriate instrument to solve this tension – a perspective which has hitherto been neglected in the doctrine. The paper argues that the freedom of expression in an online environment can be endangered through the use of machine learning for personalisation, demotion of fake news (disinformation), automated blocking of illegal content and automated enforcement of the right to be forgotten. However, the application of Article 11 of the Charter to these interferences faces insurmountable obstacles due to its personal scope of application. Given that the text of this provision prevents only interferences by public authorities and thus seems to expressly exclude those by private companies such as (social) media, it is questionable whether the Charter can prevent such interferences at all. Indeed, the Charter in principle protects the citizens only against actions of public authorities (vertical direct effect) and not against other private entities (horizontal direct effect). Despite the recent recognition of horizontal direct effect of non-discrimination, effective judicial protection and the right to paid annual leave in Egenberger and Bauer, this paper argues that a similar conclusion cannot be made for freedom of expression and information. In the absence of comprehensive EU secondary legislation on freedom of expression which could compensate for such a lack of horizontal effect, citizens seem to be left without adequate protection of their rights in cases of unjustified removal of content wrongfully labelled as disinformation or illegal. This calls for further legislative action on the level of the EU.
Download the article from SSRN at the link.
Comments