UDC 336.1
DOI: 10.51409/v.a.2023.10.03.011
Authors
O. E. Bashina
Moscow University for the Humanities, Moscow, Russia
L. V. Matraeva,
MIREA – Russian Technological University, Moscow, Russia
E. S. Vasiutina,
MIREA – Russian Technological University, Moscow, Russia
Abstract
The article discusses the features of the transformation of the «principal-agent» problem under the influence of the
introduction of technologies based on artificial intelligence. The authors show that along with the classic manifestation of
the “principal – agent” problem, in which artificial intelligence plays the role of an effective tool for effective risk reduction,
an approach is now becoming relevant in which AI becomes an internal participant in the dilemma, namely, an agent.
Algorithms embedded in AI for effective decision making can provoke new contradictions between the goal setting
of the principal (human) and the goal on which the agent is focused (AI technology), which leads to an aggravation of
this problem.
It has been proven that managers, when using AI algorithms, must understand two main limitations: firstly, AI
algorithms do not take into account any other factors than those that are embedded in them, because of this, they can
get out of hand if not applied properly. Second, AI algorithms are black boxes. Although they can predict the future with
great accuracy, they will not tell you what causes an event or why. It is concluded that managers, taking into account these
features, will be able to overcome the challenges of the “principal – agent” problem in its new interpretation.
Keywords
“principal – agent” problem, artificial intelligence, artificial intelligence with a teacher, artificial intelligence without a teacher, risk, introducing.