Artificial intelligence between the lines: Navigating economic inequality and mental health

Abstract

Economic inequality is a significant determinant of mental health, as financial insecurity and social disadvantage often intensify anxiety, depression, and psychological distress. This review explores the moderating role of artificial intelligence (AI) in this relationship, illustrating how AI can either buffer or amplify the psychological effects of economic hardship. AI-driven technologies such as predictive analytics, chatbots, and digital mental health platforms have expanded opportunities to identify and support individuals vulnerable to mental distress, particularly in resource-limited environments. These tools can mitigate inequality-related mental health risks by improving access to affordable and personalized care. However, disparities in digital access, AI literacy, and data representation can reinforce existing disadvantages, worsening psychological vulnerability among marginalized groups. By analyzing these contrasting roles, this paper emphasizes the importance of equitable access, ethical design, and inclusive implementation in ensuring that AI functions as a protective rather than divisive force. Understanding how AI moderates the connection between economic inequality and mental health is crucial for developing fair and sustainable approaches to psychological well-being in an increasingly digital world.

Key Words: Artificial intelligence; Economic inequality; Mental health; Artificial intelligence literacy; Digital access

Core Tip: Artificial intelligence moderates the relationship between economic inequality and mental health in two ways. It can alleviate psychological burdens by improving access to care, or exacerbate distress by reinforcing digital and social exclusion. Building equitable systems and ensuring inclusive artificial intelligence use are essential for realizing its protective potential.

INTRODUCTION

Economic inequality remains one of the most powerful social determinants of mental health. Individuals who experience persistent financial strain, job insecurity, or limited access to social resources are more likely to face anxiety, depression, and feelings of hopelessness[1,2]. The rapid integration of artificial intelligence (AI) into everyday life has added a new dimension to this relationship. AI now influences healthcare, employment, education, and welfare systems, and thus shapes both the causes and consequences of inequality-related stress.

While AI holds promise for reducing mental health disparities, it also introduces new risks. For instance, AI systems can identify vulnerable individuals earlier than traditional screening methods and deliver timely, personalized psychological support[3-5]. However, AI technologies often depend on digital infrastructure, data quality, and technical literacy, which remain unevenly distributed across economic groups[6,7]. This unequal access indicates that underserved individuals in need of support - low-income or digitally excluded individuals - benefit the least from AI innovations.

This review examines how AI moderates the link between economic inequality and mental health. Rather than viewing AI solely as a technological advancement, this paper interprets it as a dynamic factor that alters the strength and direction of the relationship between financial hardship and psychological well-being. The discussion focuses on two contrasting functions of AI: Its buffering role in mitigating mental distress and its exacerbating role in reinforcing social and digital divides. To avoid redundancy, key concepts such as the digital divide, AI literacy, and algorithmic bias are introduced here as interconnected mechanisms and are subsequently examined in greater depth within specific sections, each serving a distinct analytical purpose.

REVIEW PROCESSES

The present review employed a structured narrative approach to integrate current evidence on the intersections among AI, economic inequality, and mental health. The purpose was to provide a comprehensive synthesis of recent developments and ongoing challenges in this emerging field. To ensure thoroughness and clarity, multiple stages were followed in identifying, selecting, and analyzing relevant literature. An extensive search was conducted across multidisciplinary databases, including PubMed, Scopus, PsycINFO, Web of Science, and Google Scholar. The search covered publications between 2015 and 2025 to capture recent empirical findings related to AI-based mental health support, digital inclusion, and economic inequality. Combinations of terms such as “artificial intelligence”, “economic stress”, “economic strain”, “economic challenge”, “mental health”, “digital access”, “inequality”, “emotional well-being”, “psychological well-being”, and “psychological outcomes” were used.

Studies were included if they examined mental health outcomes in relation to AI-based systems, digital mental health platforms, or technology-mediated interventions that reflect socioeconomic disparities. Literature addressing ethical, accessibility, or social aspects of AI was also included to provide contextual understanding. Excluded were purely technical or engineering-oriented articles that lacked a social or psychological focus. The identified works were reviewed and organized into thematic categories that aligned with the aims of the study. These included: (1) AI as a tool for improving accessibility and early mental health detection; (2) The impact of digital exclusion and AI literacy gaps on psychological well-being; (3) Potential biases that reinforce socioeconomic divides; and (4) Policy and ethical considerations surrounding equitable AI implementation. The initial database search yielded approximately 420 records. After removing duplicate entries, titles and abstracts were screened for relevance, resulting in 96 articles selected for full-text review. Following full-text assessment based on relevance to AI applications, economic inequality, and mental health outcomes, a total of 52 studies were included in the final synthesis. Representative search strings combined terms related to AI (e.g., “artificial intelligence”, “machine learning”, “AI-based systems”) with indicators of economic challenge (e.g., “economic inequality”, “financial strain”, “low-income populations”) and mental health outcomes (e.g., “depression”, “anxiety”, “psychological distress”, “mental well-being”).

Through this process, the review sought to highlight how AI serves as a moderator - both protective and amplifying - in the association between economic inequality and mental health. The synthesis was interpretative rather than statistical, focusing on conceptual clarity and policy relevance. This approach enabled the integration of findings across health, technology, and social welfare disciplines to construct a multidimensional understanding of the topic.

AI AS A MODERATING MECHANISM

The following subsections examine how previously introduced themes - digital access, AI literacy, and algorithmic bias - operate differently depending on whether AI functions as a buffering or exacerbating moderator of the relationship between economic inequality and mental health (Table 1).

Table 1 Framework of artificial intelligence’s moderating role between economic inequality and mental health.

Key concept	Academic summary
AI as a moderating mechanism	AI operates as a dynamic factor that can either buffer or intensify the psychological effects of economic hardship
Protective function of AI	When equitably accessible, AI tools such as chatbots and predictive systems improve access to mental health support and early intervention for disadvantaged populations
Exacerbating function of AI	Unequal access, limited AI literacy, and algorithmic bias can reinforce disparities and increase psychological distress among low-income groups
Digital accessibility	Access to affordable internet and digital devices determines whether AI promotes inclusion or exclusion in mental health care
AI literacy and empowerment	AI literacy supports confident engagement with digital health tools and strengthens psychological self-management
Ethical governance	Fair, transparent, and accountable AI systems are essential to prevent harm to economically vulnerable users
Integration with human care	AI should support, not replace, human-centered mental health care by assisting assessment and monitoring
Policy and global collaboration	Cross-sector coordination is necessary to ensure equitable AI implementation and reduce mental health disparities

AI: Artificial intelligence.

AI as a buffering moderator

AI can act as a protective factor that weakens the psychological effects of economic inequality. Digital tools supported by AI make mental health resources more accessible, even in low-resource or remote areas[8,9]. Predictive systems can identify individuals at high risk of depression or anxiety based on social and behavioral indicators, enabling timely preventive interventions[3-5,10]. Chatbots and virtual therapy programs provide emotional support and self-management guidance without the need for in-person consultation, reducing the barriers created by cost, distance, or stigma[11,12]. For example, AI-based predictive systems using electronic health records and behavioral data have been applied in public and primary care settings to identify individuals at elevated risk of depression or anxiety, enabling early outreach in populations with limited access to specialist care. Similarly, chatbot-based mental health interventions have been deployed in resource-limited contexts to provide psychoeducation, symptom monitoring, and basic emotional support at low cost. Evidence from community and primary care settings suggests that such chatbot tools can achieve high user engagement and acceptability among low-income users, demonstrating their feasibility as accessible mental health resources where traditional services are scarce. AI-driven mental health applications have the capacity to deliver personalized support that adapts to a user’s emotional state, literacy level, and socioeconomic background. By offering such individualized care, AI can help restore a sense of agency and psychological control among individuals who experience economic hardship. For communities where professional mental health care is limited, AI serves as an alternative means of coping, potentially reducing the intensity of inequality-related distress. When effectively integrated into public health systems, these technologies can narrow the mental health gap between privileged and underserved populations. These applied findings illustrate how AI can function as a buffering moderator by weakening the association between economic hardship and psychological distress through improved access to timely and personalized mental health support.

AI as an exacerbating moderator

Despite its potential benefits, AI can also amplify the mental health effects of economic inequality. Disparities in internet access, device ownership, and digital literacy continue to limit participation in AI-based health systems[13,14]. Individuals with fewer financial resources may not have access to reliable networks or digital devices, which prevents them from using AI tools designed to support mental well-being[13,14]. This exclusion generates a new form of psychological vulnerability rooted in technological inequity. Moreover, the quality and fairness of AI outputs depend heavily on the data used to train them. If datasets underrepresent low-income or marginalized groups, AI systems can produce biased or inaccurate results[5,15]. For example, automated diagnostic systems may make an error in classifying distress in minority or disadvantaged populations, leading to delayed or inadequate support. The emotional consequences of such misrepresentation - feeling unseen, misjudged, or excluded by technology - can intensify the psychological burden of inequality. In institutional settings such as workplaces, education, and healthcare, AI systems can inadvertently reproduce structural disadvantages. Automated tools used for recruitment, academic evaluation, or patient triage may rely on data patterns that reflect existing social inequalities. When individuals from lower socioeconomic backgrounds are disproportionately filtered out or misclassified by these systems, they may experience feelings of unfairness, loss of agency, and heightened anxiety about their social mobility or access to care. Such experiences can erode trust in technology and institutions, adding a new dimension of psychological stress to the existing burden of economic hardship. In this context, existing evidence demonstrates how AI systems may act as an amplifying moderator, intensifying the mental health consequences of economic inequality when biased data or unequal access limits the accuracy and fairness of technological interventions.

Data privacy concerns further intensify these risks, particularly for economically disadvantaged individuals who often have limited power to control how their mental health data are collected, stored, or reused. Access to digital mental health platforms may require users to consent to extensive data-sharing practices, and individuals facing financial hardship may feel compelled to accept such conditions in exchange for care. This imbalance in bargaining power increases the likelihood of data misuse or unintended surveillance and can contribute to heightened anxiety, mistrust, and psychological distress, thereby amplifying the mental health consequences of economic inequality.

Contextual moderation and variability

The moderating influence of AI is not static but shaped by broader social and institutional contexts. In regions with strong digital infrastructure and ethical governance, AI may serve primarily as a protective mechanism that promotes early detection, preventive care, and psychological resilience[3-5]. In contrast, where digital exclusion and unregulated technological development persist, AI tends to deepen disparities and undermine mental health equity[16,17]. The same technology can therefore produce opposite outcomes depending on its accessibility, regulation, and public trust. This demonstrates that AI’s role as a moderator is relative rather than determined. It interacts with social structures, user capacities, and policy frameworks. Its impact depends on whether it complements or conflicts with efforts to reduce inequality and improve mental health services. These patterns clarify that prior empirical findings align with a moderating interpretation, in which AI either dampens or strengthens the psychological impact of economic inequality depending on contextual and institutional conditions.

POLICY AND PRACTICE IMPLICATIONS

Recognizing the moderating role of AI in the relationship between economic inequality and mental health has profound implications for both policy development and professional practice. Ensuring that AI functions as a supportive rather than amplifying moderator requires an integrative approach that bridges technology governance, social welfare policy, and public health practice. Digital equity must be treated as a public health priority. The psychological benefits of AI-based interventions depend on equal access to the underlying digital infrastructure[8,13,14]. Expanding affordable internet services, supporting device accessibility, and promoting community-based digital hubs can prevent digital exclusion from becoming a new form of mental health risk. Governments and local authorities should view digital access not as a technological privilege but as an essential condition for psychological and social participation. Without this foundation, AI’s potential to moderate inequality remains unrealized.

AI literacy programs should also be embedded into lifelong education and social service systems. A lack of understanding about AI systems can produce anxiety, mistrust, or alienation, particularly among older adults and economically disadvantaged groups[18]. For economically disadvantaged populations, efforts to improve AI literacy are most effective when embedded within familiar and trusted community-based settings. Programs delivered through public libraries, community centers, nonprofit organizations, and social service agencies can integrate basic digital skills training with hands-on demonstrations of AI-based health tools. Such approaches lower entry barriers for individuals with limited technological experience and have been shown to increase confidence, sustained engagement, and willingness to use digital mental health resources. By situating AI literacy within everyday community contexts, these initiatives help ensure that technological competence functions as a source of empowerment rather than exclusion. Structured training on how to use AI tools, evaluate their reliability, and manage digital privacy can enhance users’ sense of control and agency. Integrating AI education into school curricula, community centers, and health outreach programs can help transform digital literacy into a mental health resource that builds confidence and resilience in the face of rapid technological change.

Ethical governance and accountability mechanisms are essential for protecting mental well-being in AI deployment. Mental health algorithms, decision-support systems, and predictive analytics must be designed and monitored according to clear standards of fairness, transparency, and human oversight. Independent review boards and public ethics committees can evaluate AI applications to prevent bias, discrimination, and data misuse. Developers should be required to disclose the logic and limitations of algorithms that influence mental health assessments or service eligibility, ensuring that individuals are not harmed by opaque or inequitable technological systems.

AI must be integrated into mental health care in ways that complement rather than replace human support. While digital therapy platforms and chatbots provide cost-effective and scalable interventions, the emotional complexity of mental distress requires empathy, cultural understanding, and clinical judgment that AI alone cannot replicate. Hybrid care models - where AI tools assist clinicians by identifying risks, tracking symptoms, or managing routine tasks - can allow professionals to focus on relational and therapeutic dimensions of care. This approach ensures that technology enhances the human connection central to psychological recovery.

Cross-sector collaboration is needed to align technological innovation with social welfare objectives. Mental health professionals, policymakers, technologists, educators, and community organizations should jointly design frameworks that promote inclusion and well-being. Collaboration across different fields helps turn technological progress into solutions that are socially relevant and responsive. This cooperation ensures that advances in AI reflect the real experiences of people living with economic challenges and mental health difficulties. Governments should support international cooperation on AI ethics and public mental health, recognizing that digital inequality and its mental health consequences are global concerns requiring coordinated responses.

Collectively, these policy and practice directions underscore the need for a comprehensive vision in which technology serves as a catalyst for psychological equity. The ultimate goal is to regulate AI as well as to embed it within a social system that prioritizes empathy, fairness, and accessibility. When guided by these principles, AI can help transform the structural determinants of inequality into opportunities for empowerment and improved mental health outcomes. These policy implications are derived from the consolidated discussion of digital access, literacy, and governance presented earlier, translating conceptual insights into practical directions without restating prior analyses.

Although the included studies provide valuable insights into the intersection of AI, economic inequality, and mental health, the overall evidence base remains uneven. Many studies relied on cross-sectional designs, pilot implementations, or exploratory analyses, which limits the ability to draw causal conclusions. Longitudinal studies and rigorous evaluations examining how AI modifies the relationship between economic hardship and mental health over time were relatively scarce. In addition, evidence from low-resource settings was often based on small-scale or feasibility-focused research. These limitations highlight the need for more robust empirical investigations to strengthen the evidence supporting AI’s moderating role. Despite growing scholarly attention to AI-enabled mental health interventions, several critical gaps remain in the current literature. Empirical studies that explicitly test AI as a moderating factor between economic inequality and mental health outcomes are relatively scarce, with most existing research relying on descriptive or associative analyses. In addition, low-resource and economically marginalized settings remain underrepresented, limiting the generalizability of current findings and constraining understanding of how AI functions under conditions of limited infrastructure and institutional support. Addressing these gaps is essential for advancing both theoretical clarity and practical relevance in this field.

CONCLUSION

AI has become a defining factor in how societies experience and respond to economic inequality and its psychological consequences. Many studies suggests that AI moderates this relationship in complex and dynamic ways, capable of serving as both a bridge to inclusion and a barrier to equity[13-15]. Its role depends on the systems that govern its accessibility, the accuracy of the data on which it operates, and the degree of social readiness to use it responsibly[5,15]. When equitably distributed, AI can alleviate the mental strain associated with financial hardship by increasing access to care, enhancing self-management of emotional challenges, and providing new forms of psychosocial connection. However, when deployed unevenly, it risks magnifying distress, reinforcing exclusion, and embedding economic disparities more deeply into the psychological landscape of individuals and communities.

Future research should move beyond descriptive accounts and prioritize empirical designs that directly test the moderating role of AI in the relationship between economic inequality and mental health. Longitudinal studies are needed to examine how sustained exposure to AI-based systems influences mental health trajectories over time, particularly under conditions of economic instability. In addition, comparative research across socioeconomic and geographic contexts would help clarify how variations in digital infrastructure and access shape AI’s buffering or amplifying effects. Researchers should also place greater emphasis on participatory and community-informed approaches to ensure that the perspectives of economically disadvantaged populations are meaningfully incorporated into study design and interpretation.

A future-oriented approach should regard AI as a force that directly shapes mental health outcomes among economically disadvantaged populations, rather than viewing it merely as a technological tool. Policies promoting digital inclusion, data transparency, and ethical oversight will be critical to ensuring that AI’s moderating influence supports mental well-being rather than undermines it. As technology continues to evolve, the capacity to integrate AI into equitable and compassionate systems will determine whether it becomes a protective force that narrows mental health gaps or a disruptive one that widens them.

Existing reviews in the fields of digital mental health and AI have primarily examined the effectiveness, feasibility, or ethical risks of specific AI-based interventions. In contrast, the present review contributes to the broader scientific conversation by reframing AI as a contextual force that shapes how economic inequality translates into mental health outcomes. By emphasizing AI’s moderating role, this manuscript moves beyond intervention-focused perspectives and offers a more integrative understanding of how technological systems interact with structural disadvantage. This perspective provides a complementary lens to existing reviews and underscores the importance of situating AI within wider social and economic contexts.