This article examines algorithmic bias as a pressing legal challenge, situating the issue within the broader context of artificial intelligence (AI) governance. We employed comparative legal analysis and reviewed pertinent regulatory documents to examine how the fragmented U.S. approaches and the EU’s user-centric legal frameworks, such as the GDPR, DSA, and AI Act, address the systemic risks posed by biased algorithms. The findings underscore persistent enforcement gaps, particularly concerning opaque black-box algorithmic design, which hampers bias detection and remediation. The paper highlights how current regulatory efforts disproportionately affect marginalized communities and fail to provide effective protection across jurisdictions. It also identifies structural imbalances in legal instruments, particularly in relation to risk classification, transparency, and fairness standards. Notably, emerging regulations often lack the technical and ethical capacity for implementation. We argue that global cooperation is not only necessary but inevitable, as regional solutions alone are insufficient to govern transnational AI systems. Without harmonized international standards, algorithmic bias will continue to reproduce existing inequalities under the guise of objectivity. The article advocates for inclusive, cross-sectoral collaboration among governments, developers, and civil society to ensure the responsible development of AI and uphold fundamental rights.
New technologies have opened several risks to safety of journalists. More importantly, in the state of emergency caused by the Covid-19 outbreak, journalists and media actors have shifted their activities online more than ever, which also made them more prone to digital threats and attacks. In some regimes there are even organized intimidation campaigns against political opponents causing chilling effect and self-censorship, and jeopardizing freedom of expression in general. Hungary as a member of the European Union since 2004 and Serbia as a leading candidate to join the EU are two countries where the problems and concerns about media freedom is growing every day. The fear from the unknown during the international pandemic gave opportunity to some governments to hide their real political agendas and cover their desire for the 'good-old-fashioned' censorship. The number of countries where some kind of censorship could be found is growing every day. The authors will show two country-case-studies from Hungary and Serbia, where the leaders and the political situations are very similar and could show a (good or bad) example to other countries that would like to follow the illiberal views on media issues.
The article argues that the Digital Services Act, as part of the EU’s broader attempt to regulate intermediary services providers in a constantly growing and challenging technological, social and political environment, does not provide a final and comprehensive solution to the issue. The Digital Services Act appears inconsistent with the previous case law of European supranational judiciary forums regarding the prohibitionofgeneralmonitoringbyintermediary services providers.Infact,itprovides the Member States with vaguely worded regulatory exceptions in the event of a ban on general monitoring. However, the Digital Services Act can be seen as a legitimate and necessary attempt to enforce the regulation at the European level through Member States while at the same time giving a unique regulatory position in specific cases to the European Commission or pan-European legal bodies in general. Finally, the Digital Services Act also turns initial enforced self-regulatory attempts to regulate social harms, possibly caused by intermediary services providers, into co-regulation.
3 January 2025 in European Journal of Law and Political Science
After a brief, general introduction to AI, the present article will discuss whether AI itself has freedom of expression or whether it only entitles to tech companies that own AI. All this is relevant in the context of whether we can consider AI a separate legal actor in the context of content creation on social media, which is the second main issue of this article. Here, negative and positive cases of content moderation will be presented, i.e., whether the full spectrum of content moderation in social media can be entrusted to AI as a separate actor with respect to liability, or whether moderation requires some form of human control, direction or supervision.
The article shows snapshots from the history of fake news and terms and concepts to use in the field of information disorder (fake news, disinformation, misinformation) based on a thorough literature review. It describes that neither the analysis of only the the speaker’s side (the intention) nor the the recipient’s side (the impact) could not alone give a proper conceptual framework, but both should be used by the legislation to tackle these issues. With the viral misinformation through social media, it is clear that this poses new regulatory challenges and requires rapid-response frameworks suited to the digital age’s unique dynamics. In the European Union, the road was marked by the Code of Practice on Disinformation, and the article describes how the EU Digital Services Act (DSA) tries to tackle fake news and disinformation. The article gives some possible legal solutions also.