Contrary to the promise that AI will transform various industries, there are conflicting views on the impact of AI on firm performance. We argue that existing AI capability measures have three major limitations, limiting our understanding of the impact of AI in business. First, the definition of AI itself is still elusive in the IS and business literature. With the recognition that AI is a multifaceted problem-solving process different from traditional IT, we present a detailed AI classification scheme using various sources (e.g., PapersWithCode, HuggingFace, ACM). Second, existing measures on AI capability do not distinguish between stated strategies and actual AI implementations. To distinguish “AI washing” and actual AI capability, we collect various AI-related data sources, including AI conferences (e.g., NeurIPS, ICML, ICLR), AI software projects (GitHub repositories), inter-firm transactions related to AI adoption (FactSet), and AI strategies stated in 10-K annual reports. Third, while prior studies identified successful AI implementation factors (e.g., data integrity and intelligence augmentation) in a general context, little is known about the fit of specific AI technologies and different types of tasks. We draw on the theory of task-technology fit to construct a fine-grained AI capability measure that captures the unique characteristics of different industries. Using our newly proposed AI capability measure and a novel dataset, we will study the impact of AI on firm innovation and market reaction, contributing to the nascent literature on managing AI.
The rapid growth in e-commerce has led to a concomitant increase in consumers’ reliance on digital word-of-mouth to inform their choices. As such, there is an increasing incentive for sellers to solicit reviews for their products. Recent studies have examined the direct effect of receiving incentives or introducing incentive policy on review writing behavior. However, since incentivized reviews are often only a small proportion of the overall reviews on a platform, it is important to understand whether their presence on the platform has spillover effects on the unincentivized reviews which are often in the majority. Using the state-of-the-art language model, Bidirectional Encoder Representations from Transformers (BERT) to identify incentivized reviews, a document embedding method, Doc2Vec to create matched pairs of Amazon and non-Amazon branded products, and a natural experiment caused by a policy change on Amazon.com in October 2016, we conduct a difference-in-differences analysis to identify the spillover effects of banning incentivized reviews on unincentivized reviews. Our results suggest that there are positive spillover effects of the ban on the review sentiment, length, helpfulness, and frequency, suggesting that the policy stimulates more reviews in the short-run and more positive, lengthy, and helpful reviews in the long run. Thus, we find that the presence of incentivized reviews on the platform poisons the well of reviews for unincentivized reviews.
Kwon, Soonjae, Sung-Hyuk Park, Gene Moo Lee, Dongwon Lee. “Learning Faces to Predict Matching Probability in an Online Dating Market”. Work-in-progress.
Presentations: DS 2021, AIMLBA 2021, WITS 2021
Based on an industry collaboration
With the increasing use of online matching markets, predicting the matching probability among users is crucial for better market design. Although previous studies have constructed visual features to predict the matching probability, facial features extracted by deep learning have not been widely used. By predicting user attractiveness in an online dating market, we find that deep learning-enabled facial features can significantly enhance prediction accuracy. We also predict the attractiveness at various evaluator groups and explain their different preferences based on the theory of evolutionary psychology. Furthermore, we propose a novel method to visually interpret deep learning-enabled facial features using the latest deep learning-based generative model. Our work contributes to IS researchers utilizing facial features using deep learning and interpreting them to investigate underlying mechanisms in online matching markets. From a practical perspective, matching platforms can predict matching probability more accurately for better market design and recommender systems for maximizing the matching outcome.
Lee, Myunghwan, Gene Moo Lee, Donghyuk Shin, Sang-Pil Han (2020) “Robots Serve Humans: Does AI Robot Adoption Enhance Operational Efficiency and Customer Experience?”, Working Paper.
Presented at WITS (2020), KrAIS (2020), UBC (2021)
Research assistants: Raymond Situ, Gallant Tang
Service providers have been adopting various robotics technologies to improve operational efficiency and increase customer satisfaction. Robotics technologies bring new restaurant experiences to customers by taking orders, cooking, and serving. While the impact of industrial robots has been well documented in the literature, little is known about the impact of customer-facing service robot adoption. To fill this gap, this work-in-progress study aims to analyze the impact of service robot adoption on restaurant service quality using 4,612 restaurants and their online customer reviews. We analyzed the treated effect of robot adoption using a difference-in-differences approach with propensity score and exact matching. Estimation results show that restaurant robot adoption has a positive impact on customer satisfaction, specifically on perceived food quality and perceived value. This study provides both academic and practical implications on the emerging AI robotics techniques.
Schulte-Althoff, Matthias, Gene Moo Lee, Hannes Rothe, Robert Kauffman, Daniel Fuerstenau.“What Fuels Growth? A Comparative Analysis of the Scaling Intensity of AI Start-ups”. Under Review. [ResearchGate]
Presented at HICSS 2021 (SITES mini-track), Copenhagen Business School 2021, FU Berlin 2021, University of Cologne 2021, University of Bremen 2021, Humboldt Institute for Internet and Society 2021, University of British Columbia 2022.
AI technologies automate ever more complex tasks and promise new efficiencies for firms to provide new market offerings and grow. Economists argue that complementarities from AI innovations have not diffused widely enough to yield higher productivity yet though. We examine how firm revenue scales with labor for revenue-per-employee (RPE) and is moderated by firm-level AI investment. We compare AI start-ups, in which AI provides a competitive advantage, with digital platform and service start-ups. We use propensity score matching (PSM) to explain the scaling of start-ups and find evidence for sublinear scaling intensity for revenue as a function of labor. Surprisingly, our study suggests similar scaling intensities between AI and service start-ups, while platform start-ups produce higher scaling intensities. We show that an increase in employee counts is associated with major increases in revenue for platform start-ups, while increases were modest for service and AI start-ups. We also consider AI-enabled service start-ups that incorporate both service and AI-based business models and AI-enabled platform start-ups that combine AI and platform business models. AI-enabled service start-ups have a scaling intensity between service and AI start-ups, so they may not yet have achieved scaling benefits because AI adoption requires manual work from human experts. AI-enabled platform start-ups, in contrast, have a higher scaling intensity. Our study provides new perspectives on the role of AI as an emerging technology resource that supports economies of scale and scope for start-ups.
Identifying inter-firm relationships is critical in understanding the industry landscape. However, due to the dynamic nature of such relationships, it is challenging to capture corporate social networks in a scalable and timely manner. To address this issue, this research develops a framework to build corporate social network representations by applying natural language processing (NLP) techniques on a corpus of 10-K filings, describing the reporting firms’ perceived relationships with other firms. Our framework uses named-entity recognition (NER) to locate the corporate names in the text, topic modeling to identify types of relationships included, and BERT to predict the type of relationship described in each sentence. To show the value of the network measures created by the proposed framework, we conduct two empirical analyses to see their impacts on firm performance. The first study shows that competition relationship and in-degree measurements on all relationship types have prediction power in estimating future earnings. The second study focuses on the difference between individual perspectives in an inter-firm social network. Such a difference is measured by the direction of mentions and is an indicator of a firm’s success in network governance. Receiving more mentions from other firms is a positive signal to network governance and it shows a significant positive correlation with firm performance next year.
Park, Sungho, Gene Moo Lee, Donghyuk Shin, Sang-Pil Han. “Targeting Pre-Roll Ads using Video Analytics”, Under Reject and Resubmit, Management Science [Decision: July 2021]
Funded by Sauder Exploratory Research Grant 2020
Presented at Southern Methodist University (2020), University of Washington (2020), INFORMS (2020), AIMLBA (2020), WITS (2020), HKUST (2021), Maryland (2021), American University (2021), National University of Singapore (2021)
Research assistants: Raymond Situ, Miguel Valarao
Pre-roll video ads continue to rise at an unparalleled pace, creating new opportunities and challenges. They are more immersive than conventional banner ads and must be viewed at least partially before the content video is played. On the other hand, the prevailing skippable format of pre-roll video ads that allows viewers to skip ads after five seconds generates opportunity costs for advertisers and online platforms when the ad is skipped. Against this backdrop, we propose a novel video analytics method for improving pre-roll video ad performance by extracting multi-modal (audio, video, text) properties from both video ads and content videos using deep learning and signal processing techniques, and then analyzing their effect on video ad completion. The findings indicate that the ad-content congruence in various modalities is essential in explaining viewers’ ad completion. Specifically, visual congruence (i.e., celebrity overlap in ad and content) and textual congruence (i.e., topic similarity of ad and content) play important roles as viewers may shape ex-ante expectations of the congruence based on visual cues (i.e., thumbnail images) and previous experience (i.e., watched content clips from the same program) before watching the content video. We also discover, through predictive analyses, that video ad completion can be reliably predicted by features derived from the proposed method. Surprisingly, there is no discernible loss of predictive power when analyzing only the first five seconds of ads and content videos rather than their entire length, resulting in significant cost savings when processing large video datasets.
Previous title: On the Heterogeneity of Digital Infrastructure in Entrepreneurial Ecosystems
Digital infrastructure is the backbone on which digital startups realize business opportunities, and the homogeneity or heterogeneity of the technological base can have significant downstream impacts on business risks, inflexibilities, and growth barriers. On the nexus of digital entrepreneurship and infrastructure studies, we suggest a conception of startup digital infrastructure as organized in tech stacks; tech stacks contain individual technological elements that are combined in a single startup, while the way this is done will be inspired by shared templates within the ecosystem. Given that there is limited understanding of the heterogeneity (or homogeneity) of startup tech stacks, we use public registry datasets from StackShare and Crunchbase to identify common tech stacks of startups. Through our analysis, we identify ten commonly used startup tech stacks, which we use to measure and analyze the heterogeneity of startup tech stacks and its antecedents. OLS regression analysis shows that a startup’s technologies’ interrelatedness, its age, and investor funding are associated with the heterogeneity of startup tech stacks. The overall analysis suggests that while startups may make individual choices regarding technology usage, there could be underlying commonalities and imprinting effects across startups, exposing them to common risks in terms of their digital infrastructures. This could pose important implications for startups, investors, and society at large