Applying Reflexivity to Artificial Intelligence for Researching Marginalized Communities and Real-World Problems (CC BY-NC-ND 4.0)

2023 | HICSS | Citations: 0

Authors: Aguilar, N.; Landau, A.; Mathiyazhagan, S.; Auyeung, A.; Dillard, S.; Patton, D.

Abstract: Despite advances in artificial intelligence (AI), ethical principles have been ... Expand

Abstract: Despite advances in artificial intelligence (AI), ethical principles have been overlooked, harming marginalized communities. These flaws are due to a lack of critical insight into the complex positionality of the researcher, power dynamics between scholars and the communities being studied, and the structural impact on real-world problems when AI systems appear to be accurate but ethically fail. Reflexivity is a process that yields a better understanding of community-specific nuances, areas requiring local expertise, and the potential consequences of scholastic interventions for real-world problems (i.e., social, environmental, or socioeconomic). The paper builds on the five stages of social work reflexivity that can be applied to AI researchers and provided questions that can be asked in order to increase privacy, accountability, and fairness. We discuss the effective implementation of reflexivity in research, detail the stages of social work reflexivity and highlight key questions for AI researchers to ask throughout the research process. Collapse

Topics: artificial intelligence privacy unintended consequence

Methods: natural language processing image processing case study

A Multi-Label Classification with an Adversarial-Based Denoising Autoencoder for Medical Image Annotation

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chai, Yidong; Liu, Hongyan; Xu, Jie; Samtani, Sagar; Jiang, Yuanchun; Liu, Haoxin

Abstract: Medical image annotation aims to automatically describe the content of medical i ... Expand

Abstract: Medical image annotation aims to automatically describe the content of medical images. It helps doctors to understand the content of medical images and make better informed decisions like diagnoses. Existing methods mainly follow the approach for natural images and fail to emphasize the object abnormalities, which is the essence of medical images annotation. In light of this, we propose to transform the medical image annotation to a multi-label classification problem, where object abnormalities are focused directly. However, extant multi-label classification studies rely on arduous feature engineering, or do not solve label correlation issues well in medical images. To solve these problems, we propose a novel deep learning model where a frequent pattern mining component and an adversarial-based denoising autoencoder component are introduced. Extensive experiments are conducted on a real retinal image dataset to evaluate the performance of the proposed model. Results indicate that the proposed model significantly outperforms image captioning baselines and multi-label classification baselines. Collapse

Topics: workstation

Methods: data transformation autoencoder experiment deep learning literature study

Match Made by Humans: A Critical Enquiry into Human-Machine Configurations in Data Labelling

2023 | HICSS | Citations: 0

Authors: Chandhiramowuli, S.; Chaudhuri, B.

Abstract: We present a critical ethnographic study of data labelling conducted in a Banga ... Expand

Abstract: We present a critical ethnographic study of data labelling conducted in a Bangalore-based AI start-up. Labelled datasets are primarily produced by human data workers. We explore how humans and machines are configured together in data labelling and what are the demands placed on human workers, including on their body and cognition, while being assigned in the service of machine intelligence. We also show how these human-machine configurations sustain and reproduce the seamless functioning of apparently "autonomous" AI as a normative vision. Though labelled datasets are an indispensable prerequisite to creating ML/AI-based systems, the human labour that produces these datasets cannot be acknowledged fully if the techno-entrepreneurial vision of "self-learning" machine intelligence is to be celebrated and sustained. In pursuit of a normative position of what AI should be, we are left with a denial of how AI is actually produced now. Collapse

Topics: artificial intelligence price management organization structure outsourcing crowdsourcing

Methods: qualitative interview ethnography participant observation personal interview grounded theory

Theories: sociotechnical systems theory

Pro or Con? Introducing AI Ethics Debates in an Undergraduate MIS Course

2023 | Americas Conference on Information Systems | Citations: 0

Authors: Chen, Yu; Jensen, Scott; Roldan, Malu; Harper, Shannon

Abstract: Artificial intelligence (AI) is becoming pervasive across industries, making it ... Expand

Abstract: Artificial intelligence (AI) is becoming pervasive across industries, making it important for management information systems (MIS) students to understand the ethical issues involved. We present an active learning approach to teaching the fast-changing topic of AI ethics using a debate format. This approach was piloted in an undergraduate MIS course of 30 students. Over a five-week period, student teams were assigned to argue either the opportunities (pro) or dangers (con) viewpoint for five different AI technologies. A post-project survey indicated this format helps students gain a better understanding of the applications, opportunities, and potential misuse of AI. Students found this to be an engaging and fun way to explore the multiple dimensions of AI ethics that also required them to employ critical thinking, collaboration, research, and communication skills. We share our findings about the benefits of co-creating knowledge in the classroom using a debate format to explore an evolving topic. Collapse

Topics: responsible artificial intelligence artificial intelligence innovation management digital innovation robotic

Methods: survey personal interview image processing design methodology

Integration of Computer Vision with Analogical Reasoning for Characterizing Unknowns

2023 | HICSS | Citations: 0

Authors: Combs, K.; Bihl, T.; Ganapathy, S.

Abstract: Current state-of-the-art artificial intelligence struggles with accurate interp ... Expand

Abstract: Current state-of-the-art artificial intelligence struggles with accurate interpretation of out-of-library (OOL) objects. One method proposed remedy is analogical reasoning (AR), which utilizes abductive reasoning to draw inferences on an unfamiliar scenario given knowledge about a similar familiar scenario. Currently, applications of visual AR gravitate toward analogy-formatted image problems rather than to computer vision data sets. The Image Recognition Through Analogical Reasoning Algorithm (IRTARA) approach described herein shows how AR can be leveraged to improve computer vision in OOL situations. IRTARA produces a word-based term frequency list that characterizes the OOL object of interest. To evaluate the quality of the results of IRTARA, both quantitative and qualitative assessments are used, including a baseline to compare the automated methods with human-generated results. Fifteen OOL objects were tested using IRTARA, which showed consistent results across all three evaluation methods on the objects that performed exceptionally well or poorly overall. Collapse

Topics: floppy disk Wikipedia Python ergonomics usability

Methods: computational algorithm experiment word embedding GloVe survey

Let Artificial Intelligence Be Your Shelf Watchdog: The Impact of Intelligent Image Processing-Powered Shelf Monitoring on Product Sales

2023 | Management Information Systems Quarterly | Citations: 0

Authors: Deng, Yipu; Zheng, Jinyang; Huang, Liqiang; Kannan, Karthik

Abstract: We collaborated with a leading fast-moving consumer goods (FMCG) manufacturer t ... Expand

Abstract: We collaborated with a leading fast-moving consumer goods (FMCG) manufacturer to investigate how intelligent image processing (IIP)-based shelf monitoring aids manufacturers' shelf management by using data from a quasi-experiment and a field experiment. We discovered that such artificial intelligence (AI) assistance significantly and consistently improves product sales. Several underlying mechanisms were revealed by our quantitative and qualitative analysis. First, retailers are more likely to comply due to the greater monitoring effectiveness enabled by AI assistance. Second, the positive effect of IIP-based shelf monitoring partially persists after it is terminated, implying that human learning takes place. Third, the value of IIP-based shelf monitoring can be attributed to independent retailers rather than chain retailers. Since the degree of contract heterogeneity is the major difference between these retailers in terms of monitoring, this finding further suggests that AI is relatively more scalable when coping with more heterogeneous instances. Apart from these great benefits, we demonstrate the low marginal costs of implementing IIP-powered shelf monitoring, which indicates its long-term applicability and potential to generate incremental value. Our research contributes to several literature streams and provides managerial insights for practitioners who consider AI-assisted operational models. Collapse

Topics: artificial intelligence organizational value cost benefit analysis mobile application human AI collaboration

Methods: experimental group field experiment qualitative interview experiment quasi experiment

A Vietnamese Handwritten Text Recognition Pipeline for Tetanus Medical Records

2023 | International Conference on Information Systems | Citations: 0

Authors: Dinh, Minh; Le, Mau Toan; Bui, Triet; Mai, Minh; Tran, Long; Nguyen, Nhan; Vo, Tan Hoang

Abstract: Machine learning techniques are successful for optical character recognition ta ... Expand

Abstract: Machine learning techniques are successful for optical character recognition tasks, especially in recognizing handwriting. However, recognizing Vietnamese handwriting is challenging with the presence of extra six distinctive tonal symbols and vowels. Such a challenge is amplified given the handwriting of health workers in an emergency care setting, where staff is under constant pressure to record the well-being of patients. In this study, we aim to digitize the handwriting of Vietnamese health workers. We develop a complete handwritten text recognition pipeline that receives scanned documents, detects, and enhances the handwriting text areas of interest, transcribes the images into computer text, and finally auto-corrects invalid words and terms to achieve high accuracy. From experiments with medical documents written by 30 doctors and nurses from the Tetanus Emergency Care unit at the Hospital for Tropical Diseases, we obtain promising results of 2% and 12% for Character Error Rate and Word Error Rate, respectively. Collapse

Topics: healthcare data digital transformation database system mobile system data security

Methods: experiment image processing deep learning computational algorithm machine learning

AgriMitr: Digitalizing the Agricultural Landscape with Satellite Imaging

2023 | Communications of the Association for Information Systems | Citations: 0

Authors: Ghosh, Riju; Dhaigude, Amol

Abstract: AgriMitr was a tech-savvy start-up situated in Bengaluru, the Silicon Valley of ... Expand

Abstract: AgriMitr was a tech-savvy start-up situated in Bengaluru, the Silicon Valley of India. It was a subsidiary of KnowThyData (KTD), a Data Science and Analytics firm, a leading data science firm providing myriad solutions to many clients. AgriMitr was established by the founder and CEO of KTD, Dr Jishu Bose, to develop products using satellite imagery and data science. AgriMitr developed a mobile and web application that helped farmers and other key stakeholders understand crop health, soil fertility and other critical factors related to agriculture by machine learning algorithms using satellite images. The app was a major leap of innovation in the Indian Agri-Tech space. The app was to be used majorly by farmers based in rural India, and thus the major challenge lay in spreading awareness, adoption, and usage of the app. The app, if appropriately deployed, would be an imminent milestone in sustainable food production & consumption and can aid in getting a step forward in solving the world hunger problem, an SDG of the UN. Dr Bose was planning the strategy for AgriMitr's mass adoption and implementation with the help of Mr Madhav A. Ram. The much-awaited meeting with investors and key stakeholders was just one week away. Madhav, under the guidance of Dr Bose, had to prepare a roadmap and his game plan for the next year to bring this wave of change and large-scale digital adoption among the farmers of India. Collapse

Topics: mobile application digital literacy

Methods: machine learning computational algorithm longitudinal research survey synthesis

When Images Backfire: The Effect of Customer-Generated Images on Product Rating Dynamics

2023 | Information Systems Research | Citations: 0

Authors: Guan, Yue; Tan, Yong; Wei, Qiang; Chen, Guoqing

Abstract: Customer-generated images (CGIs) on e-commerce platforms have been widely adopte ... Expand

Abstract: Customer-generated images (CGIs) on e-commerce platforms have been widely adopted as a promotional tool to persuade customers into purchases. Despite their prevalent applications, the effect of CGIs on customer postpurchase satisfaction has not been extensively examined. This study postulates that CGIs could cause expectation disconfirmation and reduce product uncertainty for customers, therefore making their effect on subsequent product ratings complex. We leverage multiple methods and data sets to gain a better understanding of this problem and underlying mechanisms. We employ a difference-in-differences model to empirically test our hypotheses and find that CGIs lead to a decline in subsequent ratings compared with product ratings not exposed to CGIs. Further heterogeneity analyses demonstrate that high CGI review rating and high aesthetic quality exacerbate the negative effect, whereas reviewer face disclosure could alleviate the negative effect. Through cross-product analyses, we find that the negative effect is more prominent for experience goods (e.g., women’s dresses) than for search goods (e.g., lightning cables). Finally, the underlying mechanism is further validated through a laboratory experiment that shows participants experience significantly higher expectation and more negative disconfirmation in the CGI group with high review ratings, whereas uncertainty reduction effect is insignificant, which collectively explains the decline of subsequent product ratings. These findings suggest that platforms and retailers should be aware of the potential negative effect of CGIs on the rating dynamics and take appropriate measures to circumvent it. History: Ravi Bapna, Senior Editor; Gordon Burtch, Associate Editor. Funding: This work was supported in part by National Science Foundation of China [Grants 72202220, 72172070, and 71729001]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2023.1201. Collapse

Topics: online review experience good electronic commerce data breach digital good

Methods: experiment experimental group laboratory experiment computational algorithm product review

‘I could eat a horse’: The Impact of Hyperbole on Product Sales on Short Video Platforms

2023 | International Conference on Information Systems | Citations: 0

Authors: Han, Yutong; Yin, Bowen; Ren, Fei; Gu, Bin

Abstract: ... Expand

Abstract: Collapse

Topics: advertising management marketing management consumer behavior decision making Douyin

Methods: regression analysis method computational algorithm descriptive statistic natural language processing ImageNet