| Abstract [eng] |
The advancement of digital technologies and the integration of artificial intelligence (AI) systems are reshaping traditional employment relations and challenging existing models of legal liability in civil and labour law. Contemporary AI systems that operate autonomously and make decisions in contexts significantly affecting human rights raise complex questions regarding the distribution of responsibility between the user – typically the employer – and the AI developer. Due to opacity, autonomy, and the risk of inherent algorithmic bias, classical tort liability models often prove insufficient. The first part of the thesis examines the legal concept of liability and its key components: unlawful conduct, fault, damage, and causality. It distinguishes between subjective and objective liability and introduces alternative theoretical approaches, such as functional, engineering, and risk-distribution models, all of which are particularly relevant in the context of AI. The second part explores European Union and national legislation regulating the use of AI in employment contexts. It reviews the AI Act (Regulation (EU) 2024/1689), the proposal for an AI Liability Directive (COM(2022) 496 final), the General Data Protection Regulation (GDPR), and national legal instruments governing responsibility for automated decision-making. The third part addresses practical liability issues arising when algorithmic decisions cause harm to workers, such as discrimination, dismissal, or reputational damage. The influence of the 2023 Council of Europe Convention on AI and Human Rights (the Vilnius Convention) is analysed, emphasizing the necessity for national legal doctrines to extend the concept of objective liability to capture the structural characteristics of AI. The thesis concludes that the current legal framework for liability does not adequately reflect the realities of technological decision-making. A transformation of legal doctrine is therefore required – both in labour law and civil liability – towards recognising collective, distributed, and impersonal forms of liability in response to the specific nature of AI-driven systems. |
