Article (in Estonian): O.-J. Luik, K. Saluste, M. Volberg, "Kas tehisarule tuleks anda õigussubjektsus?" (Should artificial intelligence be granted legal personality?), Juridica 2026/4, pp. 331–341. Article page on Juridica (published 2026)
Co-authors: Kadi Saluste and Mats Volberg.
Research question
Liability rules assume that someone understands and controls the conduct that caused harm. With autonomous AI that assumption weakens. The article asks whether AI should still be treated solely as a product for which the manufacturer or user answers, or whether AI conduct should carry legal significance of its own.
Main arguments
The authors show that treating AI purely as a product leaves gaps, for instance where the manufacturer or user can rely on force majeure and the harm goes uncompensated. The question therefore shifts from who is liable to whether and when AI conduct should have its own legal weight. The article weighs the arguments for and against legal personality and asks whether civil law needs a third category beside natural and legal persons, a sui generis "electronic person", or whether the problems can be solved otherwise. The authors support an electronic person with limited legal capacity, but oppose extending human or civil rights to AI.
Relevance for insurers
Liability gaps feed directly into liability, product liability and cyber insurance, since an insurer pays within the insured's liability. The EU Artificial Intelligence Act regulates requirements for AI systems but does not create legal personality for AI. A fuller English discussion is in Should AI have legal personality?.