מפגש פורום בינה מלאכותית בעסקים

​

Meeting agenda:

Abstracts:

Presenter: Moshe Unger

Title: Don’t Need All Eggs in One Basket: Reconstructing Universal Embeddings of Customers from Individual-Domain Embeddings

Abstract: Although building a comprehensive view of a customer has been a long-standing goal in marketing, this challenge has not been successfully addressed in many marketing applications because fractured customer data stored across different “silos” are hard to integrate under “one roof” due to numerous technical, security and privacy, organizational, political, and regulatory issues. In this talk, I will present a method of integrating several partial customer views into one consolidated customer profile without explicitly combining domain-specific customer data. This is achieved by building the universal embedding of a customer from several partial embeddings, which is theoretically possible because of Kolmogorov’s Mapping Neural Network Existence Theorem. I will also show that the method of reconstructing the universal customer embedding can obtain a personal user representation that is close to the actual ground-truth universal embedding obtained from the integrated customer data. Moreover, I will show that the reconstructed universal embeddings perform as well as the partial and the actual universal embeddings in the recommendation and other prediction tasks for two industrial applications.

Presenter: Itai Linzen

Title: AI Can Help Counter Self-Threats

Abstract:  In recent years there has been a growing body of academic research concerning AI aversion and its adverse effects on consumers in different domains (e.g., Dietvorst et al., 2015; Leung et al., 2018; Longoni et al., 2019; Mende et al., 2019). However, it is well known that people benefit immensely from AI and its applications, both at present, and more importantly, in the future. Therefore, it is important to understand not only when people avoid AI but also when and why people will embrace it and seek its help. In this research, we study AI appreciation through the scope of compensatory consumption, the idea that consumers experiencing a threat to self (e.g., to their sense of control, financial wellbeing, power) will seek products and services with self-enhancing properties. In four studies, we show that (1) participants experiencing a threat to their sense of control show higher preference to AI relative to non-threatened participants; (2) that this effect is mediated by threatened participants’ perception that AI can help them manage their life and restore it into order; (3) that this effect depends on AI’s perceived learning capabilities; and (4) that AI restores reported sense of control among threatened participants. These studies expand the scope of human-AI interaction research and the field of compensatory consumption. In addition, it may provide valuable tools for AI designers and inform policymakers of the positive effects AI can have on human wellbeing.