To what extent will consumers accept recommendations made by artificial intelligence and to what degree will such recommendations influence purchasing decisions? The factors, as research from the Boston University and University of Virginia, are complex and one important aspect is the type of product and its purpose.
An important consideration when deciding how to incorporate artificial intelligence recommenders in retail is based upon whether consumers are focused on the functional and practical aspects of a product (what the researchers refer to as its “utilitarian value”) or on the experiential and sensory aspects of a product (what is described as the “hedonic value”).
Such research is important given that more and more companies are utilizing new advances in AI, machine learning, and natural language processing. Within retail, there is considerable focus in terms of providing recommendations to consumers.
As part of the investment in AI firms need to understand whether consumers will react in sufficient numbers to product recommendations. This rests on whether consumers will trust the “word of machine,” and what leads to situations when they resist being direct to what they should or should not like.
To reach some answers, the research group examined data from over 3,000 study participants. This enabled patterns to be considered in relation to utilitarian and hedonic aspects of a product and AI recommenders.
Central to the analysis was a theory termed the ‘word-of-machine effect’. This is the belief that AI systems are more competent than humans at dispensing advice when it comes to the utilitarian functions of a product and less effective when it comes to the hedonic qualities.
The research confirmed this, leading to the suggestion for retailers that when presented with a nudge to choose products based solely on utilitarian/functional attributes, more consumers chose AI-recommended products. However, when presented with hedonic attributes, a higher percentage of consumers elect to choose human recommenders.
As an example of the models tested, one part included the selection of a winter coat. When utilitarian attributes were presented, such as warmth and breathability, the AI options were preferred by consumers. Whereas, with hedonic attributes like fabric type, the human recommendations were preferable.
Furthermore, customers who are more likely to be satisfied with “one size fits all” recommendations are more inclined to rely on AI-systems. However, customers who desire personalized recommendations are more inclined to rely on humans.
The research appears in the Journal of Marketing, in a research paper titled “Artificial Intelligence in Utilitarian vs. Hedonic Contexts: The “Word-of-Machine” Effect.”
