Economic and Societal Impacts of Technology, Data, and Algorithms
Permanent URI for this collection
Browse
Recent Submissions
Item Genuine or Fake? Explaining Consumers’ Perception and Detection of AI-Generated Fake Reviews(2025-01-07) Fröhnel, Kim; Santelmann, Bennet; Zarnekow, RüdigerThe importance of online reviews for consumers' decision-making engages fraudsters to game the review system by writing or buying fake reviews. Fake reviews are a main threat to consumers since they are hardly distinguishable from genuine human-made reviews. Moreover, advances in generative AI like ChatGPT foster the simple creation of persuasive text, such as high-quality fake reviews. While prior studies primarily focused on automatic fake review detection, little is known about how consumers react to AI-generated fake reviews. Based on a quantitative-qualitative study with 151 consumers (906 review classifications), we found that humans cannot reliably distinguish between genuine and AI-generated fake reviews (accuracy= 53.2%). They are especially worse at detecting negative AI-generated fake reviews. Our findings extend prior research by examining consumers' ability to detect AI-generated fake reviews, identifying a set of cues they use for review classification, and investigating the cues' effectiveness for detection. Further, we derive practical implications.Item AI and Human Creativity on Content Creation: Analyzing Watermark and Detection Policies(2025-01-07) Wu, Zhenhua; Hu, Lin; Chen, Pei-YuThis paper explores the dynamic interplay between artificial intelligence (AI) and human creativity within digital content platforms, focusing on how AI influences creator incentives and consumer welfare through content quality. We develop a model to analyze how AI affects content creators' incentives and consumer welfare, focusing on AI's substitutive and complementary relationships with human skills and examining the effects of AI watermarking and detection policies. Our findings show that: first, AI benefits low-skill creators when human-input costs are low but may drive high-skill creators out of the market when costs are high, reducing content quality. Second, AI watermarking, which marks AI-generated content, helps high-skill creators stay if human-input costs are moderate, improving content quality. If AI complementarity is high and watermark removal rates are low, watermarking is more effective. On the demand side, watermarking enhances content quality and subscription satisfaction. Third, an AI detection service shows similar results. High-skill creators prefer watermarking when removal rates are high and detection when rates are low. Both policies improve content quality and consumer satisfaction, with specific benefits depending on the watermark removal rate.Item Machine Behaviorism: Exploring the Behavioral Dynamics of Large Language Models in Decision Making(2025-01-07) Ni, Yongxin; Xu, Tianqi; Li, Chunxiao; Zheng, Eric; Gu, BinLarge Language Models (LLMs) have empowered AI agents to autonomously make decisions on behalf of humans, exhibiting humanoid behavior. However, the intricate dynamics driving such behavior remain largely unknown, making it crucial to understand and guide the behavior of these LLM agents, especially in high-stakes decision environments. This paper proposes a new concept of Machine Behaviorism, unlocking the potential of LLM-based agents by providing a systematic framework to reshape the agent behavior. By tracking these agents’ actions under different decision environments, we identify AI behavioral biases when delegated with high-stakes decisions, investigate the origins of such behaviors, and propose corrective interventions accordingly. Specially, using a typical representational behavioral bias commonly found in human investors, we provide a detailed demonstration of machine behaviorism in the financial trading environment. Finally, we advocate for the establishment of machine behaviorism as an emerging discipline, as AI agents increasingly undertake autonomous decision-making and become more integrated into our daily lives.Item Care for the Mind Amid Chronic Diseases: An Interpretable AI Approach Using IoT(2025-01-07) Xie, Jiaheng; Zhao, Xiaohang; Liu, Xiang; Fang, XiaoMotion sensing is a cutting-edge area in chronic disease management. Depression, a widespread complication of chronic diseases, is neglected in those studies. We draw on medical literature to endorse depression prediction using motion sensor signals. To safeguard trust for this high-stake decision, we develop an interpretable deep learning model: Temporal Prototype Network (TempPNet). Because of the temporal feature of sensor signals and the progressive property of depression, TempPNet innovatively modifies existing prototype learning models by capturing the temporal symptom progression of depression. Our empirical results indicate TempPNet outperforms state-of-the-art models in predicting depression. We also interpret our prediction via the visualization of the depression temporal progression and its corresponding symptoms detected in the walking sensor signals. We contribute to the data science methodology with a temporal symptom progression-based prototype network. Patients, doctors, and caregivers can utilize our model on mobile devices to access patients’ depression risks in real-time.Item To Claim or Not to Claim? Impact of Owner's Business Page Claiming on Customer Evaluation(2025-01-07) Lee, Jong Youl; Lysyakov, Mikhail; Rui, HuaxiaLocal businesses increasingly use review platforms to monitor customer feedback on their business pages, often created by customers. Despite the benefits of claiming ownership of such pages, many remain unclaimed, raising the question: is there a downside to business page claiming? This study examines the impact of business page claiming on customer reviews using a unique dataset from Yelp. By leveraging the heterogeneous timing of owner business page claiming, we estimate that business page claiming lowers average customer rating. This is mainly driven by an increased chance of receiving the lowest customer ratings and a decreased chance of receiving the highest customer ratings. Moreover, customers who provide the lowest ratings tend to write lengthier reviews after business page claiming. While business page claiming may signal trustworthiness, this study cautions small businesses: claiming your business page is not costless even if it is free.Item Introduction to the Minitrack on Economic and Societal Impacts of Technology, Data, and Algorithms(2025-01-07) Ge, Yong; Zhang, John