Artificial intelligence and civil liability

August 13 , 2023 // Specialty

Published in Lau & Trends

Artificial intelligence has undergone a dizzying evolution in recent decades and we can affirm that it is present in all areas likely to generate legal relationships deserving of protection.

We can define artificial intelligence following the Resolution on Civil liability regime for artificial intelligence of the Special Commission for artificial intelligence, hereinafter AI, created by the European Parliament as "that system that is based on software or integrated into hardware devices, and that exhibits behavior that simulates intelligence, among other things, by collecting and processing data, analyzing and interpreting its environment, and acting, with some degree of autonomy, to achieve specific goals.

The proposed European Regulation of April 21, 2021 establishes in its article 3.1 the conditions that an AI system must have in order to be subject to its scope of application. These are: operating with data and inputs received from machines or humans, inferring how to achieve a given set of objectives using strategies based on machine learning or logic and knowledge, and generating results in the form of content, predictions, recommendations or decisions. that influence the environment with which it interacts.

The future regulation pivots on the principle that the person who creates, maintains, controls, exploits or operates the AI system must be responsible for the damage or harm caused by the device or the activity carried out by it.

It is relevant when configuring this responsibility for damages that AI systems, although aimed at a pre-established objective, are capable of offering autonomous and non-predetermined responses, which may, therefore, not depend exclusively on the programming of their algorithm, which will be even more evident when the development on a practical level of the so-called strong AI occurs in the future; and this allows us to conclude that there may be cases of harmful results derived from erroneous decision-making by these systems or from their autonomous learning known as machine learning without being able to say that they derive from an error in the system itself.

This current context of digital disruption requires the attention of the legislator and other legal operators who must be able to analyse, anticipating all aspects deserving of a legal response, new situations lacking any type of legislative precedent, taking into account that, due to the difficulty Derived from their originality, it will be necessary to combine the rapidity of changes in the technologies that serve as a base and their serious transversal impact on both individual and organizational human relations.