This paper examines how to support blue-collar workers in adapting to the growing digitalization of work. Recognizing that blue-collar workers typically possess lower levels of formal skills, the study highlights their heightened vulnerability to technological disruption and the challenges of transitioning to digitalized work environments. Using an exploratory case study of a cleaning company, the research extends career construction theory as a theoretical lens to understand how managerial interventions support worker adaptation. The findings identify a set of ten interventions. These interventions help blue-collar workers develop new competencies, adjust to digitally augmented workflows, and maintain employability in evolving job contexts. In contrast to most existing research that focuses on how blue-collar workers themselves adapt to the digitalization of their work, this study focuses on their managers, contributing to the research and practice for a more human-centered approach in supporting blue-collar workers' adaptation to digitalized work.
Organizations are increasingly using artificial intelligence (AI) to support employee performance evaluation, yet employees' perceptions of and responses to these systems within this high-stakes workplace context remain insufficiently understood. Drawing on social exchange, organizational justice, and trust theories, this study examines how employees' justice perceptions shape trust in AI-based performance evaluation systems and how trust in AI relates to broader employee–organization relationship factors, including trust in the organization, trust in managers, and affective commitment. It further investigates whether these relationships vary across gender and Eastern–Western regional contexts. Survey data from 291 employees across Eastern and Western countries were analyzed using partial least squares structural equation modeling and multigroup analysis. The findings show that distributive and procedural justice perceptions positively influence employee trust in AI across regional contexts, with distributive justice exerting the stronger effect. Trust in AI, in turn, positively influences affective commitment in both contexts. Gender differences were observed in the relationship between procedural justice perception and trust in AI within the Eastern sample but not within the Western sample. These findings contribute to research on organizational justice, trust, and AI in organizations, while also offering practical insights for fostering employee trust and supporting effective AI implementation in performance evaluation.
Hate speech detection in online environments poses significant challenges due to the nuanced and evolving nature of hate speech and the complexities of real-time communication. This study evaluates whether zero-shot large language models (LLMs), including Claude, Gemini 2.0, ChatGPT-o1, and ChatGPT-4o, can support the annotation of gaming-specific hate speech. We compare individual models, a multi-model collective, and TF-IDF–Logistic Regression classifier against an operational human-majority reference on 1,178 Dota 2 chat messages. We also examine whether ChatGPT-generated messages improve the traditional classifier when used exclusively as training-time augmentation, with full-archive augmentation and SMOTE as comparators. Among the automated approaches, the four-model collective achieved the highest accuracy and macro-F1 but did not uniformly outperform the strongest individual LLMs. For the class-sensitive measures, synthetic-data composition mattered more than volume. A post-hoc qualitative analysis identified recurring disagreement patterns and model-specific difficulties. Targeted augmentation achieved higher accuracy than SMOTE. The findings show that class-aware LLM generation can provide a competitive alternative to feature-space oversampling for lightweight classifiers, while zero-shot LLMs are best positioned as complementary annotation tools under human oversight.
This paper argues that ethical technology design cannot be meaningfully addressed without transforming the institutional gender regimes that shape the science, technology, engineering, and mathematics (STEM) ecosystem. Drawing on care ethics and institutional gender theory, it develops an integrated framework for examining both the normative assumptions embedded in technology and the organizational conditions under which it is produced. Through a narrative review of gender-focused technology design and a cross-case synthesis of empirical studies on care ethics in practice, the paper shows how gender stereotypes, role congruity expectations, and ideal-worker norms constrain participation, leadership, and the ethical content of design itself. It critiques dominant ethical paradigms, particularly utilitarian and deontological approaches, for marginalizing the embodied, relational, and contextual dimensions of excluded groups' lived experience. In response, the paper advances care ethics as a generative framework that must be institutionally embedded if it is to challenge technocratic neutrality and support more inclusive design and policy practices. Rather than presenting care ethics as a self-contained solution, the paper positions it as a critical lens for rethinking both technological design and institutional reform, thereby contributing to more just, accountable, and socially responsive technological futures.
Discrimination does not disappear in digital labor markets. It takes different forms. Using 27,871 U.S.-based Upwork profiles, this study examines how gender and racial earnings disparities persist within a platform framed as meritocratic. Drawing on impression management, signaling theory, and affect-as-information perspectives, we conceptualize freelancer profiles as active signal spaces rather than neutral listings of skills. Female freelancers and all minoritized racial groups earn less than their white male counterparts, even after accounting for observable indicators of productivity, reputation, and experience. These disparities appear to be shaped in part by platform-level structures, including algorithmic search ranking and badge-tier architecture. Profile-level signals of emotional valence and openness to new experiences are associated with earnings, but these associations vary across groups. For Black freelancers, each unit of expressed joy is associated with approximately 4.6 percent higher earnings, and signals of openness with an apparent elimination of their disadvantage. For Middle Eastern freelancers, openness signals are associated with an estimated 40.8 percent decrease in earnings, indicating that personality cues are not valued uniformly across racial groups. Formal education is negatively associated with earnings, suggesting a boundary condition for human capital theory in algorithmically mediated labor markets where behavioral and reputational signals carry greater weight.
