THE NORMATIVE SIGNIFICANCE OF THE PRECAUTIONARY PRINCIPLE IN ARTIFICIAL INTELLIGENCE PROBLEM
THE NORMATIVE SIGNIFICANCE OF THE PRECAUTIONARY PRINCIPLE IN ARTIFICIAL INTELLIGENCE PROBLEM

Summary/Abstract: Against the backdrop of the implementation process of the AI Act starting in February 2025 , and in light of the opposition from major industry companies to adopting Codes of Good Practice in this area, following an AI Summit with its “epicenter” in Paris, it seems that the narrative underlying the European model for AI regulation is set to evolve with new perspectives. Starting from the absolutization of risks to highlighting positive aspects, in the face of a possible paradigm shift, we find it useful to draw attention to what logic defines as „argumentum ad temperantiam”. At first glance, moderation appears to be a correct approach in relation to the ambivalent nature of new technologies. However, the validity of reasoning cannot depend on certain predefined extremes, as this would lead to a scenario where the reasoning itself would be contingent on the subjective way in which these “extremes” are defined. The appeal to moderation, as a logical fallacy, assumes that truth lies in an intermediate position between two extremes. And while we have consistently advocated for normative balance, we cannot consider such a compromise to be valid—at least not automatically, and not without a thorough evaluation of the factual basis behind each “extreme.” Admitting that a compromise is not always the correct solution and that truth does not depend on balance, but on evidence, we find the essence of the precautionary principle to be an appropriate approach to the issue, at least as a first level of syllogism. This is because, as we will highlight at the conclusion of this paper, the technical evolution of artificial intelligence, by increasing the degree of uncertainty, will ultimately lead to the de facto inapplicability of the precautionary principle.

• Details
• Contents