Artificial Intelligence Law And Digitalisation In Law

Artificial Intelligence Systems As Legal Entities - Stillbirth Or Premature Birth Of The "Electronic Person"?

The study explores the evolving legal discourse on the recognition of artificial intelligence (AI) systems as legal persons, commonly described as “electronic persons” (e-persons), within the framework of European Union law. It focuses on the EU Artificial Intelligence Regulation of 2024, which, despite its extensive scope and detailed regulatory provisions, notably refrains from engaging with the question of AI legal personhood. Central to the analysis is the distinction between “weak” AI, which underpins current applications characteristic of Industry 4.0, and the prospective “strong” AI anticipated to be a defining feature of the forthcoming Industry 5.0. The existence of weak AI and the prospect of future strong AI within an Industry 5.0 context have triggered growing concern among jurists and policymakers, giving rise to an increasingly heated debate over whether responsibility and accountability can continue to be clearly assigned and delimited in cases involving erroneous decisions or malfunctions of AI systems, and whether existing legal frameworks require adaptation in response to these developments. The study argues that, although the recognition of e-persons is theoretically and legally conceivable, it is neither necessary nor normatively justified under current legal and technological conditions, as demonstrated through a functionalist theory of legal personhood. By allocating responsibility to human actors—such as producers, operators, and users—modern legal frameworks, particularly those governing tort law, contract law, and product liability, remain adequate for addressing issues of accountability and liability arising from contemporary AI technologies. However, the anticipated development of self-learning and autonomous AI systems capable of operating without human assistance calls into question the continued suitability of these existing legal theories. Accordingly, the concept of the electronic person, while presently dispensable, may acquire substantial normative significance in the future as a mechanism for allocating liability to AI systems exhibiting a high degree of decision-making autonomy. Far from constituting a doctrinal dead end, e-personhood should therefore be regarded as a premature yet potentially critical legal concept, awaiting both technological maturity and further normative refinement.

Michael Martinek
DOI: 10.53478/TUBA.978-625-6110-66-3.ch05