Daisuke Akiba

Expanding the discussion to digital media, Putri (2026) offers an intersectional mini-review examining how AI-generated visual disinformation disproportionately impacts marginalized groups. This work highlights how algorithmic vulnerabilities exploit pre-existing social fault lines, deepening digital divides. On a global scale, Olaniyan et al. (2026) investigate how generative AI influences epistemic (in)justice among higher education students across the Global North and Global South. Their findings reveal stark geopolitical divides in how AI-mediated knowledge production is experienced, raising crucial questions about who controls the narrative in the age of synthetic intelligence.Algorithms do not operate in a vacuum; they interact dynamically with localized cultural contexts, regional economies, and societal structures. Chen and Gao (2026) quantify the digital cultural divide in China, evaluating how platform algorithms actively shape rural-urban identity politics. Their empirical work demonstrates that recommendation engines can heighten regional polarization by commodifying rural identities for urban consumption. In the domain of organizational communication, Sonni (2026) critically reviews AI-enhanced communication strategies within Southeast Asian startup ecosystems, showing how adopting westernized AI models without cultural translation can inadvertently erase regional communication nuances and marginalize local talent.Looking beyond present-day systems, Singh and Sabu (2026) present a conceptual analysis mapping discrimination in "futuristic societies" through the lens of speculative fiction. By tracing the transition from human bias to posthuman bias, they challenge us to contemplate how emerging technologies might construct entirely new classes of systemic oppression if left unchecked.Addressing these systemic challenges requires robust theoretical and practical frameworks for design, verification, and ethical deployment. Waters (2026) contributes a theoretical framework for AI testing, evaluation, verification, and validation (TEVV) focused on accessibility, arguing that true accessibility in digital health and public platforms cannot be an afterthought; it must be systematically audited across every stage of the AI development lifecycle. Finally, addressing the fundamental tension between algorithmic efficiency and human empathy, Schenk (2025) explores the permissibility of AI through the concepts of mechanical objectivity and empathic understanding, challenging the uncritical adoption of automated decision-making in contexts where efficiency must not replace human nuance and moral judgment.Taken together, these eight articles illustrate that AI is neither inherently empowering nor inherently oppressive. Rather, its impact is mediated by the socioeconomic, cultural, and political conditions under which it is developed and deployed. In other words, structural risks (e.g., algorithmic bias, epistemic injustice, digital divides, etc.) and equity-driven approaches (e.g., culturally informed models, accessible TEVV frameworks, intersectional design, etc.) jointly shape AI's consequences for racial and ethnic equity.Three cross-cutting insights emerge here. First, algorithmic harm is rarely a purely technical failure; rather, it is a structural phenomenon rooted in whose data, language, and lived experiences are treated as the default (Fields et al., 2025;Olaniyan et al., 2026). Second, equity cannot be retrofitted: culturally responsive design, accessibility auditing, and intersectional analysis must be embedded throughout the AI lifecycle (i.e., from problem formulation and data curation through validation and post-deployment monitoring) rather than appended after harm has occurred (Sonni, 2026;Waters, 2026). Third, context is constitutive: identical technologies yield divergent outcomes across regions, communities, and political economies, cautioning against the wholesale export of one-size-fits-all models across cultural boundaries (Chen and Gao, 2026;Putri, 2026;Schenk, 2025).Equally meaningful is what remains underexplored (see Singh and Sabu, 2026). This Research Topic invited work on AI in asylum processing and migration governance, on inclusive education, and on communities where identities or help-seeking behaviors are stigmatized or criminalized; yet, these domains remain underrepresented in the present collection and thus constitute urgent priorities for future scholarship, as do the intersections of race and ethnicity with religion, disability, and other identities, particularly in non-Western contexts.Building on the foundations laid here, we call on future research to: develop community-led ethical frameworks that center marginalized voices from the design phase onward; expand intersectional methodologies (i.e., quantitative, qualitative, and mixed) that capture the compounding effects of race, ethnicity, socioeconomic status, gender, and disability; implement enforceable auditing and TEVV standards for AI in high-stakes domains including healthcare, education, border control, and public policy; and pursue longitudinal, cross-national comparative studies that track equity impacts as technologies and regulatory landscapes co-evolve.For readers in digital health especially, the stakes are concrete: as AI-driven diagnostics, mental health tools, and public health platforms reach communities historically underserved by medicine, the difference between culturally informed and color-blind design may translate directly into disparities in care.We trust this Research Topic serves as both a comprehensive reference for scholars and a catalyst for actionable, equity-driven AI governance. The systems now reshaping our institutions were built by people; and, hence, we believe that they can be recalibrated and/or rebuilt, more justly, by people as well.

Recent benchmarks increasingly report that large language models (LLMs) exhibit human-like causal reasoning abilities, including counterfactual inference and intervention planning. However, many such evaluations rely on domains that are heavily represented in training data and embed strong semantic cues, raising the possibility that apparent causal competence may reflect semantic pattern recombination rather than structure-sensitive causal reasoning. Drawing on human developmental theories of causal induction, this perspective argues that genuine causal understanding requires robustness to novelty and reliance on conditional structure rather than semantic familiarity. To illustrate the testability of this claim, the paper includes a pilot demonstration using synthetic causal micro-worlds. Identical numerical evidence was presented to a LLM under two conditions: semantically meaningful variable labels and non-semantic coded labels. Across paired cases, the model reliably selected the correct causal structure when labels were meaningful, but frequently misidentified or exhibited instability in causal model selection under coded labels, despite producing locally coherent explanations. These divergences emerged most clearly in diagnostically ambiguous settings requiring suppression of misleading marginal associations. The results align with the claim that semantic scaffolding can support and stabilize apparent causal competence in LLMs without implying structure-sensitive reasoning.

This theoretical paper introduces the Verbal–Cognitive Scaffold (VCS) Model, a cognitively inclusive framework which proposes the cognitive architectures underlying computational thinking (CT). Moving beyond monolithic theories of cognition (e.g., executive-function and metacognitive control models), the VCS Model posits inner speech (InSp) as the predominant cognitive pathway supporting CT operations in neurotypical populations. Synthesizing interdisciplinary scholarship across cognitive science, computational theory, neurodiversity research, and others, this framework articulates distinct mechanisms through which InSp supports CT. The model specifies four primary pathways linking InSp to CT components: verbal working memory supporting decomposition, symbolic representation facilitating pattern recognition and abstraction, sequential processing enabling algorithmic thinking, and dialogic self-questioning enhancing debugging processes. Crucially, the model posits these verbally mediated pathways as modal rather than universal. Although non-verbal architectures are acknowledged as possible alternative routes, their precise mechanisms remain underspecified in the existing literature and, therefore, are not the focus of the current theoretical exploration. Given this context, this manuscript focuses on the well-documented verbal support provided by InSp. The VCS Model’s theoretical contributions include the following: (1) specification of nuanced cognitive support systems where distinct InSp functions selectively enable particular CT operations; (2) generation of empirically testable predictions regarding aptitude–pathway interactions in computational training and performance; and (3) compatibility with future empirical efforts to inquire into neurodivergent strategies that may diverge from verbal architectures, while acknowledging that these alternatives remain underexplored. Individual variations in InSp phenomenology are theorized to predict distinctive patterns of CT engagement. This comprehensive framework, thus, elaborates and extends existing verbal mediation theories by specifying how InSp supports and enables CT, while laying the groundwork for possible future inquiry into alternative, non-verbal cognitive pathways.

This theoretical paper introduces the Verbal-Cognitive Scaffold (VCS) Model, a cognitively inclusive framework which proposes the cognitive architectures underlying computational thinking (CT). Moving beyond monolithic theories of cognition (e.g., executive-function and metacognitive control models), the VCS Model posits inner speech (InSp) as the predominant cognitive pathway supporting CT operations in neurotypical populations. Synthesizing interdisciplinary scholarship across cognitive science, computational theory, neurodiversity research, and others, the framework articulates distinct mechanisms through which InSp supports CT. The model specifies four primary pathways linking InSp to CT components: verbal working memory supporting decomposition, symbolic representation facilitating pattern recognition and abstraction, sequential processing enabling algorithmic thinking, and dialogic self-questioning enhancing debugging processes. Crucially, the model posits these verbally mediated pathways as modal rather than universal. Alt-hough non-verbal architectures are acknowledged as possible alternative routes, their precise mechanisms remain underspecified in existing literature and, therefore, are not the focus of the current theoretical exploration. Given this context, this manuscript focuses on the well-documented verbal support provided by InSp. The VCS Model's theoretical contributions include: (1) specification of nuanced cognitive support systems where distinct InSp functions selectively enable particular CT operations; (2) generation of empirically testable predictions regarding aptitude–pathway interactions in computational training and performance; and (3) compatibility with future empirical efforts to inquire into neurodivergent strategies that may diverge from verbal architectures, while acknowledging that these alternatives remain underexplored. Individual variations in InSp phenomenology are theorized to predict distinctive patterns of CT engagement. This comprehensive framework, thus, elaborates and extends existing verbal mediation theories by specifying how InSp supports and enables CT, while laying the groundwork for possible future inquiry into alternative, non-verbal cognitive pathways.

This theoretical article examines how belongingness, defined as the sense that one’s participation is legitimate and valued, interacts with cultural value orientations to help explain persistent disparities in U.S. technology engagement, including emerging technologies, across racial and ethnic groups. While structural barriers (e.g., racism, poverty, linguistic bias, etc.) remain essential to understanding such inequity, we argue that engagement patterns in technology also reflect how different cultural communities may define and experience belongingness in relation to digital domains. Drawing on Triandis and Gelfand’s framework, and focusing specifically on educational contexts, we propose the Belongingness through Cultural Value Alignment (BCVA) model, whereby belongingness serves as a catalyst between cultural value orientations and technology engagement, with vertical collectivism deriving belongingness primarily through structured skill development and validation while horizontal collectivism focusing instead on belonging based on community integration. When technological environments value practices that are consistent with vertical collectivist norms, individuals from horizontal collectivist cultures may experience cultural misalignment not from disinterest in technology or exclusionary efforts but, instead, because dominant engagement modes conflict with their familiar frameworks for fostering a sense of belonging. By examining how cultural value orientations mediate the sense of belonging in contexts involving modern technologies, the proposed perspective offers a novel framework for understanding why access alone may have proven insufficient to address technological participation gaps, and suggests directions for creating technology spaces where individuals from a wider range of communities can experience the authentic sense of belonging.

This theoretical article examines how belongingness, defined as the sense that one’s participation is legitimate and valued, interacts with cultural value orientations to help explain persistent disparities in U.S. technology en-gagement, including emerging technologies, across racial and ethnic groups. While structural barriers (e.g., racism, poverty, linguistic bias, etc.) remain essential to understanding such inequity, we argue that engagement patterns in technology also reflect how different cultural communities may define and experience belongingness in relation to digital domains. Drawing on Triandis and Gelfand's (1998) framework, and focusing specifically on educational contexts, we propose the Belongingness through Cultural Value Alignment (BCVA) model, whereby belong-ingness serves as a catalyst between cultural value orientations and technology engagement, with vertical collec-tivism deriving belongingness primarily through structured skill development and validation while horizontal collectivism focusing instead on belonging based on community integration. When technological environments value practices that are consistent with vertical collectivist norms, individuals from horizontal collectivist cultures may experience cultural misalignment not from disinterest in technology or exclusionary efforts but, instead, because dominant engagement modes conflict with their familiar frameworks for fostering a sense of belonging. By examining how cultural value orientations mediate the sense of belonging in contexts involving modern tech-nologies, the proposed perspective offers a novel framework for understanding why access alone may have proven insufficient to address technological participation gaps, and suggests directions for creating technology spaces where individuals from a wider range of communities can experience the authentic sense of belonging.

As aging populations grow, especially in high-income nations, the burden of dementia-related care continues to escalate both economically and socially. In response, this article introduces Computational Thinking (CT) as a novel, non-pharmacological framework to support cognitive organization in older adults experiencing cognitive frailty. Though CT has gained traction in fields like PK-16 education, including special education, for enhancing executive function and problem-solving, it has not yet been applied to geriatric- or memory-care contexts. This perspective proposes that CT’s core components (i.e., decomposition, pattern recognition, abstraction, and algorithmic thinking) may align with residual cognitive strengths in mild-to-moderate dementia, offering a strengths-based, meta-cognitive scaffold for processing routines, instructions, and transitions. We provide theoretical rationale, interdisciplinary support, and concrete implementation examples to illustrate how CT can augment existing care models. By emphasizing cognitive structuring over costly service expansion, CT has the potential to complement current interventions and inform more efficient, person-centered dementia care strategies.

The growing prevalence of AI chatbots in everyday life has prompted educators to explore their potential applications in promoting student success, including support for classroom engagement and communication. This exploratory study emerged from semester-long observations of class participation apprehensions in an introductory educational psychology course, examining how chatbots might scaffold students toward active and independent classroom contribution. Four students experiencing situational participation anxiety voluntarily participated in a pilot intervention using AI chatbots as virtual peer partners. Following comprehensive training in AI use and prompt design given to the entire class, participants employed systematic consultation frameworks for managing classroom discourse trepidations. Data collection involved regular instructor meetings documenting student experiences, challenges, and developmental trajectories through qualitative analysis emphasizing contextual interpretation. While students reported general satisfaction with chatbot integration, implementation revealed three critical complexities: temporal misalignment between AI consultation and real-time discussion dynamics; feedback inflation creating disconnects between AI reassurance and classroom reception; and unintended progression from supportive scaffolding toward technological dependency. Individual outcomes varied, with some students developing independence while others increased reliance on external validation. AI-assisted participation interventions demonstrate both promise and limitations, requiring careful consideration of classroom dynamics. Effective implementation necessitates rehearsal-based rather than validation-focused applications, emphasizing human mentorship and community-centered approaches that preserve educational autonomy while leveraging technological scaffolding strategically.

Background: The COVID-19 pandemic presented unique and unprecedented challenges for international students, those studying at institutions of higher education outside of their home countries, due to their distinct circumstances and vulnerabilities. This literature review examines the multifaceted mental health burdens they experienced and highlights the need for targeted support and interventions. Methods: A rigorous search across three databases (i.e., PubMed, PsycINFO, and ERIC) yielded 50 empirical studies for inclusion in this literature review. A six-phase thematic analysis framework was employed to identify and synthesize key themes. Results: Seven prominent themes emerged: (1) academic and professional disruptions; (2) challenges navigating international student status; (3) social isolation and loneliness; (4) difficulties with living arrangements; (5) financial and food insecurity; (6) health concerns for self and loved ones; and (7) experiences of discrimination and xenophobia. Conclusions: This review highlights a range of tolls that mental health consequences took on international students, and it suggests the need for targeted interventions and support services to address these challenges. It also identifies critical research gaps, such as the need for longitudinal studies and comparative analyses with domestic students. The implications for inclusive policies and supportive environments to promote international students’ well-being are discussed.