Artificial intelligence in asset management: a study of its acceptance by investors

Abstract

The continuous improvement of AI leads to opportunities for actively managed funds, which are under increasing pressure from cheaper passively managed funds. However, existing studies have shown that the general acceptance of AI-based technology by retail investors for investment decisions is significantly lower than the availability of these technologies. Based on the literature and evaluations by experts, the most relevant attributes of funds for investors are determined. A choice-based experimental research design, using five attributes, including the type of fund management (human vs. AI) and the fund’s transparency, both not yet considered in scientific studies, is used in a survey with 170 pre-selected participants. Execution of the experiment and the subsequent analysis of the 1700 observations, established investors’ preference structures, and thus a ranking of all five attributes and their levels. The analysis uses a block-wise regression approach to assure as well the robustness of the established results. A latent class analysis indicates that only for a small group of investors (18.39%) does the presence of a human manager matter. The majority (81.61%) is indifferent about whether the fund is managed by an AI or a human. For fund managers, this indicates that they can exploit the advantages of AI solutions, but do not have to fear the avoidance of their products by investors if they are suitably transparent. The results also stress the requirement to develop efficient yet transparent asset management solutions. The results remain stable across socio-economic characteristics and can partially be extrapolated beyond the German sample.

Ähnliche Arbeiten