Disasters of subjective assessments and their models: to the question of instrumentalization of diagnostics of the "uncanny valley" effect

• Filimonov V.A.

  V. A. Filimonov S. L. Sobolev Institute of Mathematics, Siberian Branch of the Russian Academy of Sciences. Novosibirsk. Russia

• Chernyavskaya V.S.

  V. S. Chernyavskaya. Vladivostok State University of Economics and Service. Vladivostok. Russia

The research of the "uncanny valley" phenomenon is analyzed. The phenomenon manifests itself in the absence of a monotonous dependence of sympathy for the robot on the degree of its anthropomorphism. Examples of interpretations of the phenomenon in the context of different approaches are given. A number of studies of this effect as a manifestation and attitude to the anthropomorphism of robots, as problems of identity, overcoming organicity, and "prosthetics of consciousness" are considered. The difficulties and experience of measuring subjective assessments are described: psychosemantic methods, expert assessments. The article presents the variants of the analysis of subjective assessments, which are an extremely complex phenomenon, analyzes the experience of using the developed methods of research and diagnosis of the "uncanny valley" phenomenon, analyzes the shortcomings of the methods (questionnaires), and shows the limitations of the
results obtained. The phenomena of bistability of perception in the context of the "uncanny valley" effect are considered. The presence of hysteresis is stated as the cause of such phenomena. Recommendations on the organization of experiments in accordance with this hypothesis are given. The article substantiates the method of diagnosing the "uncanny valley" phenomenon using photographs of robots with different levels of human-like behavior and assessment scales and two diagnostic options, in accordance with the bifurcation point traversals. photos of real robots, toys, and movie char-acters. The photos were selected in such a way that the organizers of the experiment could classify them according to the scales of the degree of human likeness and threat, as well as place one photo in each position of the experiment matrix, which we called Matrix-25. Two variants of the diagnostic procedure are proposed: left bypass (L): first, the degree of similarity to the person decreases, then the threat increases; right bypass (R): first, the threat increases, then the degree of similarity to the person decreases.
Keywords: subjective assessments, catastrophe theory, hysteresis, reflection, organization of experiments.