14

https://doi.org/10.22364/juzk.80.14 | 142-149 | PDF

Cēloņsakarības konstatēšana mākslīgā intelekta darbībā

Establishment of Causality in the Operation of Artificial Intelligence

Ritvars Purmalis, Mg. iur.
Latvijas Universitātes Juridiskās fakultātes doktorants

Summary
While there is still a debate at European Union level as to which of the known models of civil liability should be used to determine a person’s civil liability for damage caused by an artificial intelligence-driven system, particular attention should be directed at the issues relating to the “transparency” and “explainability” of the operation of artificial intelligence-driven system. The purpose of this article is to provide a brief insight into the importance of establishing the causal relationship between the way in which the inputs (various datasets) available for the artificial intelligence-driven system are used as a basis to reach a particular conclusion within its autonomous operation, and the importance of this matter within the context of recent legal policy updates.

Atslēgvārdi: cēloņsakarība, mākslīgais intelekts, izskaidrojamība, mašīnmācība

Keywords: causality, artificial intelligence, explainability, machine-learning