Legal Status of Artificial Intelligence from Quantum-Theoretic Perspective

• Published in: BRICS law journal Vol. 10; no. 4; pp. 5 - 34
• Main Authors: Melnikova, E., Surov, I.
• Format: Journal Article
• Language: English
• Published: Publshing House V.Ема 01.01.2023
• Subjects: artificial intelligence; law; legal capacity; meaningful human control; quantum; responsibility gap; subjectness
• ISSN: 2409-9058; 2412-2343
• DOI: 10.21684/2412-2343-2023-10-4-5-34

Massive inclusion of artificial intelligence (AI) in the technosphere and electronic governments urges an update in legal regulation of these and related areas. The issue converges on the key question of whether AI can be endowed with legal personhood and capacity. Opposing views in this respect build on hardly compatible ethics and largely outdated scientific grounds with a clear perspective for deep cultural antagonisms and further fragmentation of the world. We contribute to this debate from the perspective of quantum cognitive science and show how it can resolve some of the current impasses. Our approach builds on the quantum-theoretic refinement of the concept of uncertainty into quantum and classical types: classical uncertainty denotes subjective ignorance of the present state of affairs, while quantum uncertainty accounts for individual freedom to construct the future. We show that legal capacity of intelligence, at bottom, is defined by the type of uncertainty it is capable to resolve. Natural intelligence, in particular, can resolve quantum uncertainties, generating genuine novelty and affective experience in the process. Classical AI, in contrast, is limited to algorithmic computation, bound to produce predefined results regardless of its complexity. Concepts of decision-making, subjectness, creativity, and personal meaning then are recognized as physically inapplicable to such systems. The proposed definitions of these terms complement and sharpen the criteria of legal capacity in the existing legislations, indicating that “autonomy” is essentially equivalent to “appreciation.” Classical AI then appears as fundamentally alien to subjectness and legal capacity both in civil and common laws, resolving a delicate contradiction between them. Quantum-empowered AI, in contrast, escapes this conclusion due to its access to quantum uncertainty, introducing novel challenges with respect to responsibility gaps and meaningful human control. The developed approach aligns with the present legal practice and ethical discourse, contributing to the scientifically informed development of law in technological societies.